Sogni: Learn logo
Markdown

Comfy Worker Release Notes

The public history of Sogni Comfy Worker releases. Each entry says whether to update, what changed for your worker, and whether it needs disk space, a setting, or a driver.

#Updating

  • Installer-managed workers follow the latest image. Use Upgrade worker in the Dashboard, or run worker-restart.bat on Windows or ./worker-restart.sh on Linux. The background updater waits until the worker is idle.
  • A deployment pinned to a numbered tag (for example sogni/comfy-worker-cu13:1.0.206) must move the tag by hand.
  • Do not interrupt active work. Prefer the managed upgrade or wait for an idle moment.
  • Use the CUDA 13 image. It needs an NVIDIA RTX 30-series or newer GPU with driver R580 or newer, and installer 1.0.181 and newer run only this image. The CUDA 12 worker image is retired.

dashboard.sogni.ai/fast-workers shows the version each worker runs and whether an update is available. The Windows and Linux scripts have their own installer release notes.

Running a Linux ARM64 host with a discrete Blackwell GPU? See the separate ARM64 RTX release notes. Running an NVIDIA DGX Spark? It has its own worker image and version numbers: see Comfy Worker (DGX Spark) release notes. The versions on this page do not apply to it.

The current application release is 1.0.290, rolling out to CUDA 13 workers on the 1.0.263 image. The current Docker image is 1.0.263. Workers that haven't moved to it yet run application release 1.0.261 on the 1.0.252 or 1.0.256 image.

#Version 1.0.290

Released: November 20, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • A worker that starts while the graphics card is still busy from the previous run now waits up to 30 seconds for the card instead of failing the start. On machines that are slow to release the card, updates no longer roll back for that reason.
  • If a later update still can't start because the graphics card didn't answer, the worker records that the cause was the machine, not the update.
  • When the worker pauses for a second or more, it now logs what it was doing at the time, so these pauses can be found and fixed.
  • Includes everything in 1.0.289.
  • Nothing else changes: same models, settings and disk use.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.289

Released: November 20, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • A job that needs a LoRA now gets the worker's full download speed. Model downloads running in the background pause while the job fetches its LoRA and pick up from the same byte afterwards, so the LoRA no longer misses its deadline on a busy connection. Background downloads still finish, however many LoRA jobs come in.
  • A worker that loses its connection in the middle of a job now notices within about 20 seconds instead of two minutes, then reconnects and picks the job back up where it can.
  • Cancelling a job before it starts rendering now confirms straight away, instead of leaving the job hanging for 30 seconds.
  • Workers on Windows (WSL2) with a full personal LoRA cache no longer pause progress updates every 10 seconds to check that cache.
  • Refreshing a worker's model list no longer stops its job progress updates for up to a minute, which could make the server give up on a running job (most often on Windows/WSL2 workers).
  • Nothing else changes: same models, settings and disk use.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.287

Released: November 20, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Adds support for local model packs: a model and its workflows can be installed as a signed folder in a worker's model storage instead of shipping in the worker. A worker without a pack sees no change and downloads nothing new.
  • Also includes: Krea 2 identity edits that run out of GPU memory get one more attempt, the worker log says why a network update drain was declined, an upstream MiniMax H3 crash fix and a sturdier image-to-3D remesh step.
  • No other change you will notice: same models, settings and disk use as before.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.280

Released: November 17, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Successful self-tests now preserve the history of consecutive invalid-output failures, so repeated failures still trigger recovery.
  • A worker restarting after repeated invalid outputs now stops accepting new jobs before cleanup begins.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.279

Released: November 17, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • A worker now stops accepting new jobs as soon as shutdown begins. A job still preparing its files cannot start rendering afterward.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.278

Released: November 17, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Changing model preferences now downloads all required files for the selected default model before other selected models, so it can become ready sooner.
  • LTX video-edit requests with no input media now fail immediately with a clear error.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.277

Released: November 17, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • RTX 3090 workers now render Qwen Image 2512 correctly. Before, those jobs failed on every RTX 3090.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.276

Released: November 17, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • MiniMax H3 Control jobs now run with every supported input: pose without a reference image, inpainting with a moving mask, and a prepared guide video on its own. Before, the first two were refused, and a prepared guide on its own rendered a placeholder clip.
  • A worker that starts while Sogni briefly restarts now keeps retrying to load its workflow list for up to about 2.5 minutes instead of stopping.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.275

Released: November 16, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Workers now use only operator settings. Anything else in .env or the deployment environment, such as internal tuning, log levels or custom workflow templates, is ignored, and the worker names what it ignored once in its startup log.
  • ComfyUI must run inside the worker container. A worker set to use ComfyUI on another machine now uses its own.
  • A Fast worker can no longer lower the 30 GB minimum RAM, and no worker can turn off the 20 GB minimum free disk. A machine below either one waits instead of connecting.
  • A worker that starts while Sogni briefly restarts now keeps retrying for up to 90 seconds instead of stopping.
  • Disk and settings: the dashboard no longer shows Attempt OOM recovery, Enable warmup, Unload models after warmup or Warmup delay. Out-of-memory retry is always on and warmup is always off. Rest settings outside the dashboard ranges now use the defaults.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.274

Released: November 16, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Video jobs that use LoRAs put much less strain on system memory (RAM), so workers with less RAM no longer slow down on them.
  • Disk and settings: nothing new.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.273

Released: November 16, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • On Windows workers with 24 GB cards, Krea 2 images with LoRAs now rarely fill video memory to the point where they slow to a crawl and get handed to another worker.
  • Workers you've paused now stay paused when they update to a new release. Before, an update unpaused them.
  • Disk and settings: nothing new.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.272

Released: November 16, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • A worker that briefly loses its connection now keeps the render it was working on more often: after reconnecting it waits longer for the server to take the result back.
  • A worker whose video memory takes a while to clear after a cancelled job no longer restarts itself. Windows (WSL2) machines saw this most.
  • Workers with many LoRAs installed restore their saved settings faster after reconnecting.
  • Disk and settings: nothing new.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.271

Released: November 16, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Workers on slower disks, Windows (WSL2) machines among them, no longer get disconnected over and over right after connecting while their saved settings are restored.
  • Workers set to only accept jobs where the artist has the Sensitive Content Filter on now also check the images and videos those jobs send them, and turn down sensitive ones before rendering. Other workers don't change.
  • Disk and settings: nothing new.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.270

Released: November 16, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • On Windows workers with an RTX 4090 or RTX 3090, a Krea 2 image that runs out of video memory and slows to a crawl is now stopped within about two minutes and handed to another worker, instead of after about six.
  • Disk and settings: nothing new.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.268

Released: November 16, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Your worker now reports how fast it downloads models and uploads finished work, so slow connections can be spotted and fixed.
  • The first time a worker starts on a machine, it runs a short download speed test (at most 10 seconds and about 100 MB) and keeps the result, so later starts skip it. If the test can't run, the worker starts normally.
  • Disk and settings: nothing new.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.267

Released: November 15, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • A worker that's still downloading its models no longer restarts, or counts a crash, when the network says none of its ready models can be served yet. It keeps downloading and rejoins as soon as another model is ready.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.266

Released: November 15, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • MiniMax H3 reference-to-video jobs free each decoded reference video once it has been read, so long reference jobs need less host memory.
  • Wan 2.2 Lightning image-to-video and text-to-video, and LTX 2.3 and 2.5 video-to-video jobs (canny, depth, pose and detailer), use about 1.6 GiB less host memory with the same output.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.265

Released: November 14, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • NVIDIA CMP 170HX cards get their own memory settings, so long MiniMax H3 jobs are less likely to run out of GPU memory.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.264

Released: November 14, 2026 PT ยท Update: CUDA 13 application update for workers on the 1.0.263 image.

  • Fewer MiniMax H3 Fast jobs fail while loading on 24 GB GPUs in machines with about 48 GB of system RAM.
  • Frees memory left over from the previous job when the next one starts, so back-to-back video jobs are less likely to run out of system RAM.

Download: About 240 MB, delivered automatically.

Docker images

#Version 1.0.263

Released: November 14, 2026 PT ยท Update: CUDA 13 Docker image and application update.

  • Supports RelayVSR 2x and 4x video upscaling, keeping all audio and the exact frame count.
  • Runs MiniMax H3 Prism reference-to-video jobs.
  • Renames the content-filter setting to Only accept jobs where the artist has the Sensitive Content Filter on and applies it to every job type, including audio, 3D, segmentation, background removal and preprocessing.
  • On multi-GPU machines, accepts jobs by the memory of the GPU it renders on.
  • Uses verified model files from a read-only models folder.
  • Moves a stalled model download to a fresh connection, keeping what it received.

Disk and settings: The Docker image is about 8.5 GB to download. RelayVSR needs 23 GB of VRAM and about 5.3 GB per upscale model; Prism needs 32 GB of VRAM and about 20 GB beyond the MiniMax H3 models. The content-filter setting stays off by default (advanced configuration). Filters can miss sensitive content; this setting does not guarantee that prompts, reference media, intermediate media, or outputs are free of it.

Docker images
  • CUDA 13: sogni/comfy-worker-cu13:1.0.263, digest sha256:a95a2306c5280f31bbe72290ccd8f0ed8d5e3271015151b569dba096257aeaac.
  • An earlier 1.0.263 image was withdrawn before any worker ran it.

#Version 1.0.261

Released: November 14, 2026 PT ยท Update: CUDA 13 application update.

  • Reduces host RAM use on workers with limited RAM when switching LoRA stacks.

Download: About 455 MB, delivered automatically.

Docker images

#Version 1.0.259

Released: November 14, 2026 PT ยท Update: CUDA 13 application update.

  • Uses less system memory when loading models and decoding video.
  • Recovers a crashed render engine inside the worker.
  • Improves recovery from input-download failures.

Download: About 455 MB, delivered automatically.

Docker images

#Version 1.0.258

Released: November 11, 2026 PT ยท Update: CUDA 13 application update.

  • Improves reliability when preparing reference files for a job.

Disk and settings: About 455 MB downloads automatically. No additional model weights or settings.

Docker images
  • No 1.0.258 Docker image is published; the application update runs on the 1.0.256 image.

#Version 1.0.257

Released: November 10, 2026 PT ยท Update: Required application update for B300 workers on the current CUDA 13 image.

  • Fixes startup on B300 GPUs using the current CUDA 13 image.

Disk and settings: About 455 MB downloads automatically. No additional model weights or settings.

Docker images
  • CUDA 13: no 1.0.257 image is published; the update applies on top of the 1.0.256 image.
  • CUDA 12: no 1.0.257 image is published; existing deployments keep their current image.

#Version 1.0.256

Released: November 10, 2026 PT ยท Update: CUDA 13 application update and Docker image.

  • Adds Only accept jobs with the Sensitive Content Filter on to worker settings. It applies to new image and video assignments; running jobs finish normally.
  • Excludes audio, 3D, and workflows that bypass the filter while the setting is enabled.
  • Restores the saved preference before accepting new assignments when the worker reconnects.
  • Cancelled jobs stop downloading inputs and converting media, freeing the worker sooner.
  • Expired input preparation fails before rendering, and cancellation stops a pending retry.

Disk and settings: The application update downloads about 455 MB automatically. No additional model weights. The setting is off by default. Enable it in the dashboard or set SENSITIVE_CONTENT_FILTER_ONLY=1 before the worker's first connection. See advanced configuration for setup and environment precedence. Filters can miss sensitive content; this setting does not guarantee that prompts, reference media, intermediate media, or outputs are free of it.

Docker images
  • CUDA 13: sogni/comfy-worker-cu13:1.0.256, digest sha256:87c7b56c32e8cccd1a46f06f3accd7a7f5b940cb3811d3adb8d8f6af3e74a52b.
  • CUDA 12: no 1.0.256 image is published; existing deployments keep their current image.

#Version 1.0.253

Released: November 9, 2026 PT ยท Update: Rolling out as an application update to CUDA 13 workers on the 1.0.252 image. No new Docker image.

  • Video object selection can now start from a drawn region, follow several objects in one mask, take selections on more than one frame, and find a lost object again.
  • A stalled render during an update now recovers without restarting the worker.

Disk and settings: About 450 MB downloads automatically. No new model weights or settings.

Docker images
  • CUDA 13: no 1.0.253 image is published; the update applies on top of the 1.0.252 image.
  • CUDA 12: no 1.0.253 image is published; existing deployments keep their current image.

#Version 1.0.252

Released: November 7, 2026 PT ยท Update: Available for qualified CUDA 13 workers. CUDA 12 is held on its existing release.

  • Adds optional object-selection masks that follow objects through video clips.
  • Improves MiniMax H3 enlargement and video export.
  • Supports returning the final video frame for follow-on edits.
  • Recovers personal LoRA downloads that stop making progress.

Disk and settings: Video selection is optional and needs about 2 GB of model storage on a local CUDA 13 worker with at least 16 GB of GPU VRAM. Existing worker settings are unchanged.

Docker images
  • CUDA 13 tag โ€” sha256:278999e47b6edd907acbc72d6c0d9d7925d7eca0dcf81981d8aafb0d3ac74be7
  • CUDA 12: no 1.0.252 image is published; existing deployments keep their current image.

#Version 1.0.251

Released: November 3, 2026 PT ยท Update: Rolling out to qualified CUDA 13 workers. RTX PRO 6000 workers stay on 1.0.249 pending qualification. Existing CUDA 12 deployments stay on 1.0.231.

  • Improves Sensitive Content Filter detection for images and video.
  • Recovers stalled render processes more reliably.
  • Improves memory handling during image and video jobs.
  • Makes media delivery more reliable on difficult network connections.

Disk and settings: About 100 MB of additional filter weights download automatically. No settings changes are needed.

Docker images
  • CUDA 12 tag โ€” sha256:f8fb5533d8ddafbdf904de483983e1065d2791aa126daae0a066f2db7dc2382e
  • CUDA 13 tag โ€” sha256:8c00bf51073beaf380662dcc590922563c90331bea2ae05a340da3a0f23d1dfc

#Version 1.0.249

Released: October 31, 2026 PT ยท Update: Recommended for CUDA 13 workers; required for the qualified 12 GB NVIDIA support below.

  • Adds tested RTX 3060 and RTX 5070 support for selected image, speech, background-removal and mask jobs on 12 GB cards.
  • Speeds up Krea 2 image jobs with LoRAs when the model is released between jobs.
  • Frees temporary memory sooner during MOSS image generation.

Disk and settings: Automatic worker download; new 12 GB workers download the models for their selected jobs.

Docker images
  • CUDA 12 tag โ€” sha256:886823898f9f03c8886308aa74a2c0dac9c391b907cb2c43bde23ee65f8ec692
  • CUDA 13 tag โ€” sha256:7aa0f9d869388b59476d4d73f5ef67a667b653b815f3b5dafcf93f11975f06ce

#Version 1.0.248

Released: October 28, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Image previews consistently show the current render.
  • Delayed preview updates stay with the generation that produced them.

Disk and settings: Small automatic download; no additional model weights or worker settings.

Docker images
  • CUDA 12 tag โ€” sha256:653de0d0a8c387537d41ddd5ff4f73567fbf7d21ace573f47719e3edc531c441
  • CUDA 13 tag โ€” sha256:d67030216347cfe55d1bcae58c15ae43ec106cda45eb3a20d3bdd0d00bac29d9

#Version 1.0.247

Released: October 27, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Wan Animate jobs release finished frames, masks and video data as soon as the job no longer needs them, so long Animate videos use much less system memory.
  • Wan 2.2 jobs use less system memory and return it after the model unloads, so Wan video now runs on workers with 32 GB of RAM.

Disk and settings: Small automatic download; no additional model weights or worker settings.

Docker images
  • CUDA 12 tag โ€” sha256:d767db7fc0fc9a0a43ea7d8d9badc46569bc4227401a045512b0c0c422e7fe8d
  • CUDA 13 tag โ€” sha256:2e718e09d8bc353830c3b0e76cddb279da907998854dd4adb97cba0a112d8c30

#Version 1.0.246

Released: October 27, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Downloads large model files faster on fast connections by fetching several parts of each file at once. Slow connections download at the same speed as before.
  • An interrupted download still resumes where it stopped, and every file is still checked before the worker uses it.

Disk and settings: Small automatic download; no additional model weights or worker settings.

Docker images
  • CUDA 12 tag โ€” sha256:20a7ce7241a9930dbeda9ea45ce6d2b7887091a3a8cd3270e7d29b9682757b73
  • CUDA 13 tag โ€” sha256:871823c0850a5f45e82056b5ce11fe88cd0e10f76952ee8cc057b6106d896bf3

#Version 1.0.245

Released: October 26, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Supports optional reference-video compression for one-stage MiniMax H3 Standard, Balanced, LightX2V Turbo and FastH3 jobs.
  • Allows tested Balanced reference-video outputs up to eight seconds and Standard outputs up to 5.17 seconds at 1344ร—768 on a 32 GB RTX 5090. Existing jobs keep their normal behavior.

Disk and settings: Small automatic download; no additional model weights or worker settings.

Docker images
  • CUDA 12 tag โ€” sha256:12f5d5991558781a1e1de76c880ef4d87f9887a0b5073334073301a5e090140b
  • CUDA 13 tag โ€” sha256:39093152fd0fef14f9c61862989e52e538ecc706a60c866fa473acbca43944fc

#Version 1.0.244

Released: October 24, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Improves image checks and checks finished videos when the Sensitive Content Filter is on.
  • Fixes color shifts in generated videos and MiniMax H3 extensions.

Disk and settings: Small automatic download; no configuration changes.

Docker images
  • CUDA 12 tag โ€” sha256:98e9d6d83ecef7b7cf38981121a8b04f37019fa1be47c75a99ccf5d1444dc969
  • CUDA 13 tag โ€” sha256:ec4fffc9a8c58a27022cd7e25606aa1c1b97bb4e40be48c50af2bd69ece000fe

#Version 1.0.243

Released: October 23, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Keeps taking jobs when Docker Desktop on Windows (WSL2) cannot lock more memory for fast GPU transfers. Earlier workers restarted instead, and some hosts then downloaded their models again.
  • Locks at most 45% of system memory for fast GPU transfers under WSL2, the same limit as on Windows. Earlier workers could try to lock 95% of it.

Disk and settings: No new downloads or settings.

Docker images
  • CUDA 12 tag โ€” sha256:01e92b9ac232127917fe5611b1546b7fba207e1445163cc0baa1e2e901be4c23
  • CUDA 13 tag โ€” sha256:c462ad40e644d696a73841a2297b69a352996f8097f31b538bd2b1e46f72a792

#Version 1.0.242

Released: October 23, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Keeps taking jobs when an update has to wait its turn. Earlier workers could stop taking jobs for up to an hour after their update was postponed.
  • Keeps the sound continuous where a MiniMax H3 extension joins the original video. Earlier extensions could fall silent for about 40 ms at the join.
  • Fits MiniMax H3 reference audio shorter than 2 seconds or longer than 15 seconds instead of failing the render.
  • Uses LoRAs a worker already has when its models folder cannot be written. Earlier workers failed those jobs.

Disk and settings: No new downloads or settings.

Docker images
  • CUDA 12 tag โ€” sha256:9c44d3bda962f196e3ec8e0ea9ac75df68bc618569cfb0a7b858ae4be66eb305
  • CUDA 13 tag โ€” sha256:36749cb1a380bda8992574eb9a978e0d5cbf091131d8109dc558e90cc6979151

#Version 1.0.241

Released: October 22, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Keeps every reference photo whole in Qwen Image Edit 2511 and its Lightning version, which cropped references to the output's shape and could cut off a face.
  • Sizes image edit references to the output instead of cropping them, at most 2 MiB each.
  • Updates a worker as soon as its current job finishes. Earlier workers also waited for background model downloads, which could keep a worker out of work for up to an hour per update.

Disk and settings: No new downloads or settings.

Docker images
  • CUDA 12 tag โ€” sha256:dcd21c9b3ddc66c2202b6c20d33de84d27abd3c345c733532328d5515bbd6be0
  • CUDA 13 tag โ€” sha256:21d220dbe9b81871319e638e050b45783c5fe778073c20c14a4b1ca61d84b837

#Version 1.0.240

Released: October 21, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Extends MiniMax H3 videos with two-stage FastH3 and three or more LoRAs, which earlier workers could not build.
  • Fixes MiniMax H3 Reference to Video failing when a reference video with an odd width or height had to be retimed.
  • Keeps MiniMax H3 Reference to Video at 1080p and 2K working with LoRAs on every tier, including FastH3 Reference to Video.
  • Applies MiniMax H3 LoRAs that train the output layer to FastH3 Reference to Video, which refused or partly skipped them before.

Disk and settings: No new downloads or settings.

Docker images
  • CUDA 12 tag โ€” sha256:081e2a291e64543663620571c1688cfdeed5c4ec470dc9f277930ea500ac83fc
  • CUDA 13 tag โ€” sha256:24b8e80c3b612bccbdef4625a4645bb1f619e3db311835dd07b311145425e2af

#Version 1.0.239

Released: October 21, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Adds FastH3 Reference to Video on CUDA 13 workers: a fast MiniMax H3 model that makes video from reference pictures, clips and sound.
  • Gives long video renders enough time for their full size and length, so slow GPUs no longer stop long Wan 2.2 and MiniMax H3 clips as timed out while they are still rendering, and still stops a Wan 2.2 render that has truly stalled.
  • Recovers from a stuck render, or from a cancelled job that leaves the GPU busy, by restarting only the render engine, so the worker keeps its downloaded models and is back in seconds.
  • Applies Krea 2 LoRAs exported by fal's trainer, which the worker previously could not use.
  • Releases pinned system memory correctly after memory pinning switches itself off, so that memory is no longer held for the rest of the run.

Disk and settings: FastH3 Reference to Video downloads about 22 GB on CUDA 13 workers with at least 32 GB VRAM. No new settings.

Docker images
  • CUDA 12 tag โ€” sha256:8c3f0c1d7ec0aa8faf0e9fc055ff53a13c8b185cfbad8df5622222cba3eed811
  • CUDA 13 tag โ€” sha256:4174650cf80f15e25d4355e1a5b7935f12f0d7598c8f31acd360cc942bcf248f

#Version 1.0.238

Released: October 20, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Supports MiniMax H3 video extensions: continues a video from where it ends, carrying its last second of motion and sound into the new footage.
  • Scales very large Reference to Video pictures down before loading them, so oversized photos no longer cost extra memory or time.

Disk and settings: No new downloads or settings.

Docker images
  • CUDA 12 tag โ€” sha256:603f038109b7760468e9eaf7fd92fdbcac9234361b07b3e18c5e8b727d7c9509
  • CUDA 13 tag โ€” sha256:d6d27c0fccf7f2a12fdba48bc1fe19b3ade1062f6dec6c179b19d7bb17a0dd4b

#Version 1.0.237

Released: October 19, 2026 PT ยท Update: Recommended for CUDA 13 workers.

  • Supports Mystic X v5 and the Action Speed Slider for MiniMax H3, including negative and higher LoRA strengths.
  • Preserves negative strengths for MOSS Image LoRAs instead of turning their effect off.
  • Adds support for optional MOSS composition controls.
  • Avoids false job timeouts while large models load or unload.

Disk and settings: No new base-model downloads or settings; selected LoRAs download when requested.

Docker images
  • CUDA 12 tag โ€” sha256:f5a709ec22a4447b0d5f9f9c6c4d4fb48d3eba0c37ded2e10c152cea720327ed
  • CUDA 13 tag โ€” sha256:29251f00e6dfec638ac91d308d394e72d9d44f921733e40fd1e882d3c8ce38f2

#Version 1.0.236

Released: October 18, 2026 PT ยท Update: Recommended for CUDA 13 workers. Replaces the withdrawn 1.0.235.

  • Adds MOSS-Image SFT, an uncensored text-to-image model, with a standard preset, an eight-step preset and Personal LoRAs on compatible workers.
  • Fixes startup when a selected model uses an adapter stored in a subfolder.
  • LTX video edits accept silent source videos and retain audio from sound-bearing inputs.

Disk and settings: MOSS downloads about 17 GB on CUDA 13 workers with at least 24 GB VRAM. No new settings.

Application update 1.0.236-runtime.1 โ€” withdrawn October 18, 2026. Managed workers remain on 1.0.236.

Docker images
  • CUDA 12 tag โ€” sha256:7b2eedb4f2618751066bd41930dd84c226c5ee1fbfa478a88f2f5cc0bcd592e4
  • CUDA 13 tag โ€” sha256:9dc15030d54fa39839cf390f24c97502ef0639cba4ec0ed4cc202c48a51afdf4

#Version 1.0.235

Released: October 18, 2026 PT ยท Update: Withdrawn. Use 1.0.236 or the current recommended release.

  • Withdrawn because the worker could exit during startup. Use 1.0.236.

Disk and settings: MOSS downloads about 17 GB on CUDA 13 workers with at least 24 GB VRAM. No new settings.

Docker images
  • CUDA 12 tag โ€” sha256:3d65f75b98be91530863e526f0d346a1591ea6293a6ac4dc72e37b4b3c53a235
  • CUDA 13 tag โ€” sha256:c77d9d1f4d668e2c744f67f8e57752915f6f4928b765e7ff905276bbf3d47d10

#Version 1.0.234

Released: October 16, 2026 PT ยท Update: Rolling out gradually; no action needed. Replaces 1.0.232.

  • Adds FLUX 3 Image Klein, Black Forest Labs' fast open image model: pictures up to 4K, edits with up to ten reference photos, and artists' own LoRAs.
  • Image jobs that start from an existing image now run only the steps their strength calls for, so light touch-ups finish several times faster.
  • A repeat of a job the worker already ran now renders again instead of returning the earlier, unpaid result.
  • MiniMax video, Qwen Image and music jobs use less temporary memory.
  • Mask jobs on sideways phone photos no longer fail with a size mismatch.
  • A worker stuck restarting after a crash now recovers within about 10 seconds.
  • A worker over its model-folder size limit still checks the models it already has.

Disk and settings: a new base image: one image download about the size of the current one, then later updates are small again. CUDA 13 workers with 16 GB or more also download FLUX 3 Image Klein, about 11 GB; its 4K pictures go only to cards with 30 GB or more. No new settings.

Docker images
  • CUDA 12 tag โ€” sha256:f2b77867cd703bbbc44cb069b6dfc8b1034ce3dce3cd93eb6aee5e55113fb587
  • CUDA 13 tag โ€” sha256:0fec55aff8823f4926ab6ded53c278ce8e319f8b6ff0afa2b729f4b7c7c6a559

#Version 1.0.233

Released: October 15, 2026 PT ยท Update: Superseded. Update to the current release.

  • Withdrawn: some image jobs that start from an existing image failed. Its changes, with that fix, ship in 1.0.234.
Docker images
  • CUDA 12 tag โ€” sha256:d1ace25b8dbd5ecc1294c36e8863b47926a128e1de086dca5a927dd8310172db
  • CUDA 13 tag โ€” sha256:6d4c7d98ce8873514f72971e0213644749c9a204dd9bf965b5611f86f0c3c9b3

#Version 1.0.232

Released: October 5, 2026 PT ยท Update: Superseded by 1.0.233, which carries all of its changes; it reached only a few workers. No action needed.

  • Short speech results, such as a one-word line, no longer fail as a corrupt file.
  • LTX 2.3 videos at 4K, and long LTX 2.3 videos at high resolution, no longer fail before the first frame.
  • Image-to-image jobs at a very high strength with many steps no longer fail.
  • A worker that started offline from its built-in release now moves to the current release once it reaches Sogni again.

Disk and settings: a new base image, so a worker moves to it with one image download about the size of the current image; later updates are small again. No new settings.

Docker images
  • CUDA 12 tag โ€” sha256:4861256ffdaf14dee9f0970afb55fda2df1952e4d06193ac06a1c509b60ecb26
  • CUDA 13 tag โ€” sha256:18da43d445cc31bbfc62eb9f98c3b9d5a589b0b658bf4648ff1df06e6c167c56

#Version 1.0.231

Released: September 30, 2026 PT ยท Update: Recommended for all workers. Replaces 1.0.230.

  • Workers that follow the latest image switch once to a self-updating image: later versions arrive as a download of about 230 MB and apply between jobs, with no new Docker image to pull.
  • A slow or dropped connection during that download now resumes where it stopped instead of starting over.
  • MiniMax video jobs no longer fail at the final decode on cards that keep part of the model in system memory.
  • Finished images are now scored by the image check in the background; delivery never waits for it.
  • Multi-GPU installs created before installer 1.0.179 need one installer update before they can switch: run ./worker-start.sh in a terminal and accept the update. Until then the worker's log says so.

Disk and settings: one image download when a worker switches, smaller than the current image; later updates download only about 230 MB. No new settings.

Docker images
  • CUDA 12 tag โ€” sha256:134829497ad6b3a514636021b7a665545de1a72c6367438e3e98f01284285869
  • CUDA 13 tag โ€” sha256:3d3445c566ff1d80b993aadff0e55fb210b202d9797376743b26ea3b3a792f6d

#Version 1.0.230

Released: September 29, 2026 PT ยท Update: Recommended for all workers; strongly recommended for MiniMax video workers. Replaces 1.0.229.

  • MiniMax first-and-last-frame videos made with a LoRA now use the same attention setting as image-to-video, which keeps fast motion from smearing.
  • The check that screens finished images is more accurate and about three times faster. It uses a new model that the worker downloads once, in the background, when it starts; renders are never held up while it downloads.
  • The worker image is about 300 MB smaller.

Disk and settings: one new download of about 104 MB (the image-check model, stored under models/safety on the shared volume); no new settings.

Docker images
  • CUDA 12 tag โ€” sha256:6b915b75759d934d8aab22c01387e5e9dee86abeed2522b215110c4918894ac4
  • CUDA 13 tag โ€” sha256:120a75ee9fc4f26e77b09cc9bfaac2c2b796bafaed499ede3c41b3655cfd6902

#Version 1.0.229

Released: September 28, 2026 PT ยท Update: Recommended for all workers; strongly recommended for MiniMax video workers. Replaces 1.0.228.

  • MiniMax image-to-video at 2K without LoRAs now uses the same attention setting as LoRA jobs, which keeps fast motion from smearing and finishes a little faster.
  • A render that comes out broken (invalid values or a blank video) now fails and is retried on another worker instead of being delivered.
  • A worker whose startup stalls now exits so it is restarted, instead of staying down.

Disk and settings: no new downloads and no new settings.

Docker images
  • CUDA 12 tag โ€” sha256:c0fda42bfa15255a274cc06ee8feaebb268685d5110ca552a027deb4c3c915b2
  • CUDA 13 tag โ€” sha256:bd39fbd97b63f6997243774566f24aedf0135748187cc5b61fce7ef7dadf6be9

#Version 1.0.228

Released: September 27, 2026 PT ยท Update: Recommended for all workers; strongly recommended for MiniMax video workers. Replaces 1.0.227.

  • MiniMax videos with LoRAs at 2K now apply the LoRAs to the first pass only, so the upscale stays sharp and finishes a little faster.
  • MiniMax videos with LoRAs keep the attention setting that stops smearing in fast motion.

Disk and settings: no new downloads and no new settings.

Docker images
  • CUDA 12 tag โ€” sha256:0890e791d9e7c07f6a242ffe83893b19e4f1820ea76c38d9ebfd7e921980dbf2
  • CUDA 13 tag โ€” sha256:e394cdc688105b3ea3670870a54b9c7f09680ab4efd80b33566c5f63bbba43c0

#Version 1.0.227

Released: September 26, 2026 PT ยท Update: Recommended for all workers; strongly recommended for MiniMax video workers on machines with about 40 GB of system memory or less. Replaces 1.0.226.

  • Starting a new MiniMax video prompt no longer gets the worker stopped for running out of system memory on machines with about 32 to 40 GB of RAM.
  • When system memory is short, the worker now frees a model it is setting aside instead of copying it into memory first, so switching models no longer briefly needs room for both.
  • A worker no longer holds earlier results in memory while it starts a new job.

Disk and settings: no new downloads and no new settings.

Docker images
  • CUDA 12 tag โ€” sha256:6bcd9534b0f2c1bdff0b9ab02be7758eee6f69d8da552636e3ac27f589baf6df
  • CUDA 13 tag โ€” sha256:98e3c967dd05d1702492b7fce83a2f5036a735f76873425736bdc2c5f0c32db2

#Version 1.0.226

Released: September 26, 2026 PT ยท Update: Recommended for all workers; strongly recommended for machines with about 32 GB of system memory and for MiniMax video workers. Replaces 1.0.225.

  • Krea 2 jobs with LoRAs no longer run up to three times slower right after another model on machines with about 32 GB of memory.
  • Cancelling a job no longer restarts workers on machines with limited memory, and a cancelled MiniMax video now stops right away.
  • MiniMax videos with LoRAs render about 30% faster on GPUs with 80 GB or more of memory, with the same output.
  • Workers can now serve MiniMax videos that use intermediate keyframes.
  • A worker whose container software was changed after installation now stops taking jobs until it is recreated from the published image.

Disk and settings: no new downloads and no new settings.

Docker images
  • CUDA 12 tag โ€” sha256:48a64a53e47c65ed67a26f767acfdb1a039451218c07a829fe40dc6b7a452bf8
  • CUDA 13 tag โ€” sha256:9c701ae86a3bfdb4843c0c6f10c34abc38783cebdfd60ca8fb4c6dc3db1d20aa

#Version 1.0.225

Released: September 25, 2026 PT ยท Update: Recommended for all workers; required for workers serving MiniMax video with LoRAs on GPUs with 80 GB or more. Replaces 1.0.224.

  • MiniMax videos that use LoRAs are sharp again on GPUs with 80 GB or more of memory; those renders take somewhat longer than before.
  • MiniMax videos no longer fail on machines where a package install inside the container replaced a required component with an older version.
  • Jobs that use LoRAs no longer occasionally fail to start while another job finishes downloading the same LoRA, a regression in 1.0.224.
  • Large finished videos no longer fail to upload on slower connections: the upload time limit now grows with the file size.
  • Upscaled videos are saved at higher quality.

Disk and settings: no new downloads and no new settings.

Docker images
  • CUDA 12 tag โ€” sha256:197976e695561df3db162f23d2a78308267e754f776d0499dbf603fa10d7d0b1
  • CUDA 13 tag โ€” sha256:fd2b860e2d371b0ac88a24982ab962c664e26c3e69409906d18abddb46fa2068

#Version 1.0.224

Released: September 24, 2026 PT ยท Update: Required for workers serving Krea 2; recommended for all workers. Replaces 1.0.223.

  • Krea 2 renders that use LoRAs no longer come back blank, a regression in 1.0.223.
  • A GPU that stays too hot to take work while idle now stops the worker with a clear fault instead of leaving it online without jobs; after the restart the worker waits for the card to cool before it rejoins.
  • Workers delete leftover job files, warmup renders and interrupted downloads, including at startup, so disk use no longer creeps up between restarts.
  • A crash no longer writes a multi-gigabyte crash dump to the host's disk.

Disk and settings: no new downloads. New setting: Guard fault timeout (minutes), default 10, sets how long an idle GPU may stay above its resume temperature before the worker stops.

Docker images
  • CUDA 12 tag โ€” sha256:e286409413fd50809015c3a6c393c548f1610e1b4299181b7aa9d62c9a045ac8
  • CUDA 13 tag โ€” sha256:a3ddcded090e6ce1c496231d61efd8b8bc721b36c5df0c79fdaa5607b87c0f1f

#Version 1.0.223

Released: September 23, 2026 PT ยท Update: Required for workers serving LTX or WAN Animate; recommended for MiniMax video workers.

  • MiniMax videos finish decoding faster and use less GPU memory during decoding.
  • MiniMax videos no longer show the circular rainbow-grid artifact corrected in this release.
  • LTX output video decoding uses less memory, preventing the reproduced long-video memory crash.
  • WAN Animate fixes a known CUDA 12 processing crash and uses less temporary memory for longer videos.
  • Image jobs respect output-format and metadata settings more reliably, and cancelled jobs stop publishing late results.

Disk and settings: MiniMax workers download a 3.2 GB replacement and remove the previous 5.2 GB file, reclaiming about 2 GB; no new settings.

Docker images
  • CUDA 12 tag โ€” sha256:19dd82f4a4216fc2ed6f97ce4148989687f6861dfd6faba141bc22579d0a9ccd
  • CUDA 13 tag โ€” sha256:2809d3b73b60930f70b9360922d4fec55d83532c7a1b58ce4ccb3fab1baab36d

#Version 1.0.222

Released: September 21, 2026 PT ยท Update: Recommended for all workers, especially workers serving WAN Animate Move and Replace.

  • WAN Animate Move and Replace now report realistic initial completion estimates, improving customer countdowns and queue wait estimates.
  • Longer WAN Animate jobs retain their rendering allowance, so the corrected estimate does not cause them to be stopped too early.
  • Disk and settings: nothing new for workers already on 1.0.221.
Docker images
  • CUDA 12 tag โ€” sha256:c38a12e58c4552932489038aca9f23272b7398c3131f148e47853f043f19b43d
  • CUDA 13 tag โ€” sha256:211b929ba1cb38640b39d29f8301e054aed8ad5e23c1bdfead8c68cf8924fc58

#Version 1.0.221

Released: September 21, 2026 PT ยท Update: Recommended for all workers. MiniMax H3 workers should update.

  • MiniMax H3 FastH3 videos made from a starting image, or from a first and last image, no longer show bright lamps, bulbs and ceiling lights wobbling or undulating, at normal size and in 2K.
  • Turning on the mature-theme model opt-out while the worker runs now stops those downloads straight away, instead of finishing the pack first.
  • Disk and settings: nothing new for workers already on 1.0.220.
Docker images
  • CUDA 12 tag โ€” sha256:6a8ce56e1aa4698c996fb61e5e8e1f56da7c510fae87a2cfb4e326047e1c546e
  • CUDA 13 tag โ€” sha256:0c2e4c677f2a2c2a9415b7d5f96aba93e92efee9e7f73142beb439595c587df5

#Version 1.0.220

Released: September 18, 2026 PT ยท Update: Recommended for all workers. MiniMax H3 workers need it to serve Reference-to-Video in 2K (and 1080p).

  • Reference-to-Video in 2K. MiniMax H3 two-stage reference-to-video now delivers 2688ร—1536 (and 1920ร—1088) from up to 9 reference images, 3 reference videos, and 3 audio clips, and every two-stage reference-to-video clip length renders on 32 GB cards.
  • 1080p and 2K reference-to-video jobs go only to CUDA 13 workers that have the FastH3 model and at least 60 GB of host RAM; 768p jobs keep the 1.0.218 rules.
  • Disk and settings: nothing new for workers already on 1.0.218.
  • 1.0.219 was published with the same changes but never recommended; 1.0.220 carries its fix. Move any 1.0.219 worker to 1.0.220.
Docker images
  • CUDA 12 tag โ€” sha256:0cc81b5b08edb2b37e50b505dbd31b20ba679f115fff33bc8d3c2c7f223e636b
  • CUDA 13 tag โ€” sha256:10fada77063bf25836c738854c8818f15ff09d8418a837e3bc17b873a1889b4d

#Version 1.0.218

Released: September 17, 2026 PT ยท Update: Recommended for all workers. MiniMax H3 workers need it for the new two-stage reference-to-video jobs.

  • Oversized uploads no longer crash the worker: reference videos and images are shrunk to what the job needs before rendering, which ends the out-of-memory restarts that video-driven jobs (animate, video-to-video, reference-to-video) caused on many hosts.
  • MiniMax H3 two-stage output no longer shows the warm glowing patch that could appear in small areas such as clothing.
  • The same FastH3 job now renders identically whether it is the first job after startup or follows another, so results are reproducible from the same seed.
  • Two-stage reference-to-video jobs go only to 1.0.218 or newer.
  • On 24 GB and 32 GB cards, Standard and Balanced MiniMax H3 jobs now go only to hosts with at least 39 GB of RAM; hosts with less kept crashing on them.
  • Disk and settings: nothing new for workers already on 1.0.217.
Docker images
  • CUDA 12 tag โ€” sha256:c465fcc5a24d07ec5b1f9e0a2d1278d0fb37c66fc32e6f639d65bec8655a4a8d
  • CUDA 13 tag โ€” sha256:d61ee715b8b0eaa1663f3149076f02c651e21fed856b64a7dd3ed71d3434c5ab

#Version 1.0.217

Released: September 14, 2026 PT ยท Update: Recommended for all workers. MiniMax H3 and Pixal3D workers need it for the new jobs below.

  • MiniMax H3 workers can take video jobs driven by an uploaded audio track, from a photo, from first and last frames, or from audio alone; these go only to 1.0.217 or newer.
  • Pixal3D workers can build a 3D model from up to four photos of an object; these jobs go only to 1.0.217 or newer.
  • MiniMax H3 2K output no longer shows the faint specks and corner lines 1.0.216 left along tile seams.
  • Disk and settings: a worker with the Pixal3D pack downloads one more 5.2 GB model on restart. No new settings.
Docker images
  • CUDA 12 tag โ€” sha256:3897b6c99643195a40fc866398ed91c10cc13d84ccd18695c3437893f93cdf2d
  • CUDA 13 tag โ€” sha256:42f289b1ab8fc5ffc1152ae87173ccd4bcbe42aad705bf38b0996abc32fdd9d2

#Version 1.0.216

Released: September 13, 2026 PT ยท Update: Recommended for all workers.

  • MiniMax H3 2K output has far fewer streaks along tile seams.
  • A Fast worker with less than 31 GB of RAM now stays disconnected and reports itself unhealthy, instead of reconnecting over and over. Relaxed workers are unchanged.
  • Disk and settings: nothing new for workers already on 1.0.215.
Docker images
  • CUDA 12 tag โ€” sha256:52737ab9bb4baac02c3fe191c0dc4eafa2ad27859b1f40a387a7be48a89e6695
  • CUDA 13 tag โ€” sha256:1fd254828464832d127c445a29ac326ed94bb501b4b82737df39e59b9bc7c2da

#Version 1.0.215

Released: September 12, 2026 PT ยท Update: Recommended for all workers. Replaces 1.0.214, which was never made current.

  • Renders that stack several personal LoRAs go only to workers on this version or newer.
  • Future changes to personal LoRA rules (strengths, trigger phrases, stack size) no longer need a worker update.
  • When a model download would run out of disk, the worker deletes idle personal LoRAs to make room instead of failing.
  • Jobs that need a specific model file go only to workers that already hold it.
  • Disk and settings: a MiniMax H3 worker that has not restarted since September 12 downloads one more 659 MB file on restart. No new settings.
Docker images
  • CUDA 12 tag โ€” sha256:761c5ef79051e77ab46677e8950f6353d281e3e5f8a2020ae0c049fc214580b5
  • CUDA 13 tag โ€” sha256:d061897cffbd776829369133721c1c1b5a2d7b3ab9abeb4b91e7def59f369a2d

#Version 1.0.214

Released: September 11, 2026 PT ยท Update: Superseded by 1.0.215 before it became current. Update to 1.0.215 instead.

  • Prepared workers for the full personal LoRA strength range. Its changes ship in 1.0.215.
Docker images
  • CUDA 12 tag โ€” sha256:69774d05e86a99c49a78894dca8d8c406bf6c417587d5c2462896007466fb2e3
  • CUDA 13 tag โ€” sha256:645ee48bb695fdd638a682f5b648f5a0c4b9daa864b906c1ee940e162f8b1d94

#Version 1.0.213

Released: September 11, 2026 PT ยท Update: Recommended for all workers. Required for MiniMax H3 workers that will serve Balanced jobs with a custom video shift.

  • MiniMax H3 Balanced runs one recipe in every mode and accepts a custom video shift. Jobs with a custom shift go only to 1.0.213 or newer.
  • Disk and settings: frees about 2 GB by removing a superseded MiniMax H3 LoRA at startup. A few H3 workers download a 780 MB adapter they did not already hold. No new settings.
Docker images
  • CUDA 12 tag โ€” sha256:9c933961dbf8689b168c5aaa154753a6bedc3c552f227619d9112e909507cdb3
  • CUDA 13 tag โ€” sha256:84f1a38b654d5d8652aff438c97a36676beee5cc8d3f373eb5bacebe019e5afd

#Version 1.0.212

Released: September 11, 2026 PT ยท Update: Recommended for all workers. Required for personal LoRA jobs, FlashVSR jobs with non-default options, and LTX audio-to-video jobs.

  • Workers can now serve personal LoRAs that customers upload and Sogni approves. Files that do not match the model are refused cleanly and the job is retried elsewhere.
  • LTX audio-to-video now animates to the uploaded soundtrack.
  • FlashVSR video upscaling accepts detail, speed and seed options. Requests that use them go only to 1.0.212 or newer.
  • MiniMax H3 gains a 2K output option. Peak GPU memory stays within the base job's.
  • A missing license file no longer blocks a workflow.
  • Disk and settings: new setting Personal LoRA cache (GiB) (PERSONAL_LORA_CACHE_GB), default 25 GiB, adjustable from the Dashboard. Set it to 0 to opt out of personal LoRA jobs. MiniMax H3 workers download a 690 MB upscaler file on their next model scan.
Docker images
  • CUDA 12 tag โ€” sha256:f67093202d0b5f5e50444ccc94a5d698e40a6372a3b0258980d504196ef58bcc
  • CUDA 13 tag โ€” sha256:b58a7d3baea9c153943809ef6f199bfa3c72b22c192bc72c35a2b7e3183ee603

#Version 1.0.211

Released: September 10, 2026 PT ยท Update: Recommended for all workers. Required for FlashVSR video upscaling.

  • New job type: FlashVSR upscales finished videos to 1080p or 1440p. Served by default on NVIDIA cards with at least 23 GB of VRAM and an Ampere or newer GPU. Other workers skip it and its download.
  • Disk and settings: eligible workers download about 6.4 GB on their next model scan. No new settings.
Docker images
  • CUDA 12 tag โ€” sha256:a8d105bd2d52e7dce07b903391f8764be14fa815aa759e24c103a453cc56af35
  • CUDA 13 tag โ€” sha256:015969ec7abf0ba263ad4b6e12a7a0c04c8394caed6f009cf1950a8d86f7e96c

#Version 1.0.210

Released: September 8, 2026 PT ยท Update: Recommended for all workers. Required for standalone BiRefNet background removal.

  • Background removal no longer fails on simple subjects. Those failures were also taking healthy workers out of background-removal work for an hour at a time.
  • Pixal3D keeps only its prompt-free reconstruction path.
  • Disk and settings: nothing new.
Docker images
  • CUDA 12 tag โ€” sha256:f36081ae51d2feb30b2a0d764d5603837b43c44443a799cfd7941d9d4daa8cc5
  • CUDA 13 tag โ€” sha256:98136e715d6f05867798d4696f7cf7a9c6db031db5957e7c9280bb0b27d45e3d

#Version 1.0.209

Released: September 8, 2026 PT ยท Update: Recommended for all workers. Required for Qwen3-TTS, standalone BiRefNet background removal, or Pixal3D. Includes the 1.0.208 MiniMax H3 fix.

  • Qwen3-TTS adds three opt-in speech workflows on CUDA 13 workers with at least 24 GB of VRAM.
  • Standalone BiRefNet background removal now delivers its transparent PNG correctly.
  • Pixal3D handles portrait inputs correctly. It remains CUDA 13 only, local ComfyUI only, 30 GB of VRAM or more.
  • Disk and settings: nothing downloads unless you select the workflow. Qwen3-TTS is about 12.7 GB, BiRefNet about 450 MB, Pixal3D about 13.4 GB.
Docker images
  • CUDA 12 tag โ€” sha256:07190b17d506f38e281fff80a97b6f96df9378b5ee10382a8110c4161eedee1b
  • CUDA 13 tag โ€” sha256:442708efb13c29296198f3adc2a3adc792945f34cc0f86e7cd94a74e6e444d28

#Version 1.0.208

Released: September 8, 2026 PT ยท Update: Superseded by 1.0.209, which carries the same fix. Update directly to 1.0.209.

  • MiniMax H3 LightX2V Turbo returns to its earlier acceleration LoRA, which gives better motion and fewer artifacts. LightX2V Turbo jobs go only to 1.0.208 or newer.
  • Pixal3D no longer needs a text prompt.
  • Disk and settings: MiniMax H3 workers swap one 1.8 GB file for another of the same size. No new settings.
Docker images
  • CUDA 12 tag โ€” sha256:714a136df6fcb1b9b8c8408729789dd7fc38ce69ff56997ee70445389dc5c32a
  • CUDA 13 tag โ€” sha256:849e22937d2f1bc939969cb77a6dcb05717e7ea60406c2897ce0c3fe2c509dbd

#Version 1.0.207

Released: September 8, 2026 PT (beta) ยท Update: Evaluation only. latest stayed on 1.0.206. Pin 1.0.207 or the beta tag to try it.

  • New opt-in workflow: Pixal3D turns one image into a textured 3D model. Offered only to CUDA 13 workers with at least 30 GB of VRAM whose ComfyUI runs on the same machine. Select it under Workflows this worker serves to download about 13 GB.
  • A worker opted into a pack with duplicate file names no longer runs out of work and restarts.
  • Z-Image Turbo accepts up to 12 steps.
  • SAM 3 image segmentation no longer fails on some results.
  • Disk and settings: nothing new unless you opt into Pixal3D.
Docker images
  • CUDA 12 tag โ€” sha256:dac3ac1754067be3e33bacfd6dd16199cfdaa3ad8c957a3dc883991cc2196c4d
  • CUDA 13 tag โ€” sha256:1fd3cbc348fab9868b07965a2ef333a486a19ba932e869fe59147a66bf50022f

#Version 1.0.206

Released: September 7, 2026 PT ยท Update: Recommended for all workers, especially MiniMax H3 workers. Pinned deployments should move to 1.0.206.

  • MiniMax H3 Turbo returns to its earlier attention settings after reports of artifacts. Longer clips return to their earlier render time.
  • Interrupted model downloads recover more reliably.
  • Disk and settings: nothing new.
Docker images
  • CUDA 12 tag โ€” sha256:72f8c4a73da0236a94d536256eddeb2f8bce4a0a01f6e6ff59526319d838bb4e
  • CUDA 13 tag โ€” sha256:51819c8237faa2acacc027107dd2f8efa69efd17b2bb4f9a127f0c878750cf4a

#Version 1.0.205

Released: September 6, 2026 PT ยท Update: Recommended for all workers. Pinned deployments should move directly to 1.0.205; do not pin 1.0.204.

  • FastH3 renders with one user LoRA are faster on 32 GB and larger cards. One sample went from 160 to 106 seconds.
  • New setting Excluded LoRAs (EXCLUDED_LORA_IDS) keeps specific LoRAs from being assigned or downloaded, and removes a file only they used once the worker is idle.
  • Opt-in workflow packs are selected with the existing Workflows this worker serves setting.
  • The Windows installer reports the RAM actually available to Docker and can update .wslconfig for you when run interactively. CUDA 13 selection now checks every GPU's compute capability.
  • Disk and settings: nothing downloads because of this update.
Docker images
  • CUDA 12 tag โ€” sha256:4de517c4400f855b8587bd6d26cfa87531c3f3fb14358c6c7a4cee761f32774e
  • CUDA 13 tag โ€” sha256:12d6d93aac1779a678dbf8739d772d0e58d5644f6a6453776594735453ee3d27

#Version 1.0.204

Released: September 6, 2026 PT ยท Update: Superseded by 1.0.205 before it was recommended. Do not pin it.

  • Published during an interrupted release. Its changes ship in 1.0.205.
Docker images
  • CUDA 12 tag โ€” sha256:af17e6a35bca432962862bc9afe4c1db7d4b6b9cbf68d7bca7fc687b0c95d354
  • CUDA 13 tag โ€” sha256:8eb18018303fbe33a5a6e2e56f6917e5ae7e1a3b2d2a25ed2307b5327bf87e36

#Version 1.0.203

Released: September 5, 2026 PT ยท Update: Recommended for all workers. Important for 1.0.193 and 1.0.194, which can no longer serve Dark Beast Krea 2. Carries everything in 1.0.201 and 1.0.202.

  • MiniMax H3 Balanced: only reference-to-video keeps the community adapter. The other Balanced modes return to the LightX2V LoRA and to their 1.0.199 speed.
  • Workers stop offering models whose files were retired from the catalog, and report a clear error if a file is missing, instead of failing job after job. Sogni stops sending that model to the worker.
  • Corrupt model files repair themselves.
  • Long MiniMax H3 Turbo clips render about 17% faster with less flicker.
  • Disk and settings: re-downloads a 1.8 GB MiniMax H3 LoRA if a previous release removed it. No new settings.
Docker images
  • CUDA 12 tag โ€” sha256:a96f5fb50f8fe986de046aeb15f8fdce04edac1765cb73b2a7d1dc38ba6fdc83
  • CUDA 13 tag โ€” sha256:776a6cf2b17b531018ca4af900898f62c21171c58d7b37b24dca8ddbdd728d50

#Version 1.0.202

Released: September 5, 2026 PT ยท Update: Superseded by 1.0.203 before it was recommended. Do not pin it.

  • A startup check withheld the ACE-Step 1.5 workflows. Fixed in 1.0.203, which carries the rest of its changes.
Docker images
  • CUDA 12 tag โ€” sha256:2b66ca9c7b532f4bd30f3a415928430b2414bdfe99b1b48a7dd31fe78cb4bcd3
  • CUDA 13 tag โ€” sha256:bae8c912054d3cbe3aa36cf7f9fabcb280e0eb466fdceb56b138107eca00e048

#Version 1.0.201

Released: September 5, 2026 PT ยท Update: Superseded by 1.0.203 before it was recommended. Do not pin it.

  • Made long MiniMax H3 Turbo clips faster and steadier. Ships in 1.0.203.
Docker images
  • CUDA 12 tag โ€” sha256:4fb10f888cba3193e5b03c79185a12df11101de3f6aef148963831f659cdeb1a
  • CUDA 13 tag โ€” sha256:c07af21ba7dd25a6ac73cc976e5dd527ba0f77f5ddb9ea0b376301d230ca3079

#Version 1.0.200

Released: September 4, 2026 PT ยท Update: Recommended for MiniMax H3 workers, and needed to keep serving H3 Balanced.

  • Every MiniMax H3 Balanced mode moves to the community adapter artists preferred. Balanced renders take about a third longer at the default size.
  • Disk and settings: removes two 1.8 GB LoRA files and downloads a 744 MB adapter on the next start. No new settings.
Docker images
  • CUDA 12 tag โ€” sha256:9af9a39db63203dd195593b7d8f761cf413171bf00eeaf3fed2d61672bc32ba7
  • CUDA 13 tag โ€” sha256:d192d539b1d3b04a4821785dfaa891a7268270848b58b1c510d537cd7423a7da

#Version 1.0.199

Released: September 4, 2026 PT ยท Update: Recommended for all workers. Important on RTX 4090 and RTX 5080 hosts.

  • Model switches are reported with their real duration. Sogni now sends cold jobs to the workers that switch fastest and asks slow switchers to switch less often.
  • RTX 4090 workers get 2 GB more VRAM for models. RTX 5080 workers get their own memory profile.
  • Disk and settings: nothing new.
Docker images
  • CUDA 12 tag โ€” sha256:39327215f8df32de89284eab73070b43500be351885acc46df2e4370b39d3e77
  • CUDA 13 tag โ€” sha256:ec9ad9517ba235d6d63c935937541238f5e309f878fa2f8f0cb2ef506af8b98e

#Version 1.0.198

Released: September 4, 2026 PT ยท Update: Recommended for all workers. Apple Silicon hosts should take it.

  • Image renders finish about two seconds faster when the artist has the Sensitive Content Filter off. What gets withheld is unchanged.
  • Mac workers: quantized models no longer crawl. NVIDIA hosts are unaffected.
  • Disk and settings: nothing new.
Docker images
  • CUDA 12 tag โ€” sha256:9c258c114b66da5fb1df952bc657a007ab9943b6ae2559fa9c0100ebf1f7c8ad
  • CUDA 13 tag โ€” sha256:e8f7328daf396fb71534419709a4f5010953af2992236353bfff7c742a4eef86

#Version 1.0.197

Released: September 4, 2026 PT ยท Update: Recommended for all workers. Important on 24 GB and larger cards.

  • A worker no longer drops its connection, and the job it was holding, while loading a large model.
  • Workers now stay on one model family for longer, so the faster LoRA renders from 1.0.196 stick.
  • Chroma image-to-image works again.
  • Mac workers no longer accept extra ComfyUI start-up arguments from the environment.
  • Disk and settings: nothing new.
Docker images
  • CUDA 12 tag โ€” sha256:43c4373d3b98336050dc1a23893459928493fa33084247e46d0b3a40ab9cebb4
  • CUDA 13 tag โ€” sha256:d7b7e2d0ea6adebbd85ba0f3b94fc651f4c41d8d17a08cb69e2c8fc0e937354b

#Version 1.0.196

Released: September 4, 2026 PT ยท Update: Recommended for all workers. Worth taking promptly on 24 GB cards.

  • LoRA renders on quantized models are much faster on 24 GB cards. On an RTX 4090 one test went from 36 seconds to under 9.
  • RTX 3090 and RTX 3080 workers stop advertising Z-Image and Qwen Image Edit, which returned corrupt output on those cards. Everything else is unchanged and the worker is not faulty.
  • A CUDA 13 worker on a driver older than R580 now holds with a clear log line instead of restarting in a loop. Use the CUDA 12 image on those hosts.
  • Disk and settings: nothing new.
Docker images
  • CUDA 12 tag โ€” sha256:14ccc19f924708f4715ab33db1c1aad19294a04fce8e7d011fd4328da9ade3c3
  • CUDA 13 tag โ€” sha256:d9b2bc30caedd314c657665a9cfa55c22641e507d268a5a5572d04b5d519061d

#Version 1.0.195

Released: September 3, 2026 PT ยท Update: Recommended for all workers. latest moved here directly from 1.0.193.

  • Dark Beast Krea 2 moves to the 3.0 checkpoint, which fixes blown-out skin and waxy textures. Renders are slower on the CUDA 12 track than on CUDA 13.
  • MiniMax H3 reference-to-video Balanced is faster at every clip length.
  • LTX-2 video-to-video no longer exhausts host RAM on long uploads.
  • Time estimates after a model switch are accurate.
  • Disk and settings: new setting EXCLUDED_WORKFLOW_IDS excludes specific models without giving up automatic model choice, and frees the disk only they used. Downloads a 13.2 GB Dark Beast file and removes the old one.
Docker images
  • CUDA 12 tag โ€” sha256:84df1bbd379c7e8336a1f1e09ba238223c8884bacf0045a5fbada6a413acd899
  • CUDA 13 tag โ€” sha256:ae7abf2109afad8da934a48df66835693ea63c773d49ccdf5af203f71b26e136

#Version 1.0.194

Released: September 3, 2026 PT ยท Update: Superseded by 1.0.195, which carries its changes.

  • A worker can now check at startup that its machine has enough disk and RAM, and stop if it does not.
  • Custom LoRAs on every FastH3 mode now run on 24 GB CUDA 13 cards, including the RTX 4090 and RTX 3090.
  • Disk and settings: optional MIN_VIABLE_DISK_GB, MIN_VIABLE_RAM_GB and HEALTH_PROBE_PORT. Nothing changes without them.
Docker images
  • CUDA 12 tag โ€” sha256:83a94651c63425a33eaeb18894f976fa2f7df357ac5a6cd244c5a8ff3b10be1e
  • CUDA 13 tag โ€” sha256:0e098204b40ce75891eb719549ebf5e2baada5152a97dd8e42d75d2a305a630a

#Version 1.0.193

Released: September 1, 2026 PT ยท Update: Required for MiniMax H3 Turbo and FastH3 Turbo workers. Recommended for all.

  • Fixes the visual artifacts from 1.0.191 and 1.0.192.
  • FastH3 Turbo is a new option about twice as fast as LightX2V Turbo. It runs on 24 GB CUDA 13 cards, including the RTX 4090 and RTX 3090. Custom LoRAs need 32 GB. Not offered on the CUDA 12 track.
  • Disk and settings: nothing new.
Docker images
  • CUDA 12 tag โ€” sha256:603a777b0af678efed10749db869f3df4e0adb82130fd355f2d5260b9a971d70
  • CUDA 13 tag โ€” sha256:da3bd23849efdd72781ed6b4438b914db41c808a75ce36f1651da08372d27a83

#Version 1.0.192

Released: September 1, 2026 PT ยท Update: Recommended for all workers. Required for new MiniMax H3 Turbo workers.

  • Large model downloads resume after a network interruption instead of starting over, so fresh workers enter service reliably.
Docker images
  • CUDA 12 tag โ€” sha256:e45b8b339a1f7aa5f36f557bd05bece9ebf793300b12b311628db16749cdb597
  • CUDA 13 tag โ€” sha256:2e68ffe377f085c6d3ed3377c54479258319d458908f2186a278814602db9233

#Version 1.0.191

Released: September 1, 2026 PT ยท Update: Superseded by 1.0.193. Do not use it for MiniMax H3 Turbo.

  • Its faster MiniMax H3 Turbo path produced visual artifacts. 1.0.193 corrects it.
Docker images
  • CUDA 12 tag โ€” sha256:5c3ac045f9c2e08087aceb9d79ec048ed574378ee43d7443cadd348b72a711aa
  • CUDA 13 tag โ€” sha256:d9fc046983bdc5b19efb4d61f81bb345fe732b06198d56da959603645dbb196b

#Version 1.0.190

Released: August 31, 2026 PT ยท Update: Required for new or ephemeral workers on 1.0.189. Recommended for all.

  • Fixes a startup regression that could leave a new worker idle instead of downloading its models and taking jobs.
  • Restores installer-managed upgrades on Windows startup, and fixes live Dashboard settings updates.
Docker images
  • CUDA 12 tag โ€” sha256:dbdf752445a72a373d827125b8f0d568ea3b9cd0e31128236b857220e32868a2
  • CUDA 13 tag โ€” sha256:0d845cf45ef003fe2ce3eeb6958a29928422a84a71c38898f974eaee44357563

#Version 1.0.189

Released: August 31, 2026 PT ยท Update: Recommended for all workers.

  • Hardware care: the worker pauses new jobs when the GPU reaches about 189ยฐF and resumes near 174ยฐF, and takes a one-minute rest after 30 minutes of actual rendering. It never interrupts an active job. Both limits are adjustable from the Dashboard, and a natural idle gap already counts as the rest.
  • Supported worker settings can be managed from the Dashboard.
  • Disk and settings: the render-time rest and thermal settings are new. No downloads.
Docker images
  • CUDA 12 tag โ€” sha256:575cba17575d798f4b1711e07e390c85a3c1904c6e9ec80a1f29f85f7c480e2c
  • CUDA 13 tag โ€” sha256:ec386a03cc5a5f73f172e8f5ac08de12c0654bc1089f14c273ef1f417e022323

#Version 1.0.188

Released: August 30, 2026 PT ยท Update: Superseded by 1.0.189. If you are on 1.0.186 or 1.0.187, update.

  • Turns off an over-cautious check that made capable GPUs decline jobs they could complete.
Docker images
  • CUDA 12 tag โ€” sha256:3ee74c773865a38c7d8111a655eed0c8470fb98f748f7919803aeb2872195d99
  • CUDA 13 tag โ€” sha256:4c2a9ecf57df1b212b5883991bfb35a96d998457fc459a3a4b4d457350010295

#Version 1.0.187

Released: August 30, 2026 PT ยท Update: Superseded. Update to the current release.

  • First correction for workers declining valid jobs. Still too cautious; fixed in 1.0.188.
Docker images
  • CUDA 12 tag โ€” sha256:b4d2fe4e5cde09b48cb54f77ccc36a8e5bfa496eefeca84fb2d6439b9a5e33a1
  • CUDA 13 tag โ€” sha256:fb36bff3cce6c22cd76e836cc49f30c7a5e5e44882a4673b8bc40b1a40ed7150

#Version 1.0.186

Released: August 29, 2026 PT ยท Update: Superseded. Update to the current release.

  • Better memory handling and recovery, especially on Windows, but one new check declined some valid jobs. Corrected in 1.0.187 and 1.0.188.
Docker images
  • CUDA 12 tag โ€” sha256:f9559f6512d4f0c47d97f701eb66f97d299106e6fa22e4962017d472a855e806
  • CUDA 13 tag โ€” sha256:813aec94e0dcb5f6808f1392af85c81d1c78d623ec1a72c8dbf5b518c5f24edd

#Version 1.0.185

Released: August 28, 2026 PT ยท Update: Superseded by 1.0.189.

  • Refreshed the MiniMax H3 Balanced acceleration files. Replaced files were cleaned up automatically.
Docker images
  • CUDA 12 tag โ€” sha256:456087cd912af74a6a32a90892930772bb9f5614f786076a5a030d294f02315a
  • CUDA 13 tag โ€” sha256:28e32ec972e846979239996b302088bb1f68e5c3c6a31df07c1026342e7b802e

#Version 1.0.184

Released: August 28, 2026 PT ยท Update: Superseded by 1.0.189.

  • Broad reliability update for MiniMax H3: previews, recovery, monitoring and compatibility.
Docker images
  • CUDA 12 tag โ€” sha256:449232604d84d1d9284c1191c9484c90690e0f88db37b097bc5fbb93ea788a62
  • CUDA 13 tag โ€” sha256:757d2ae0bc60ef5b2a051404d25c657ab602c0e2013a7dd04c99a18edcf896ae

#Version 1.0.183

Released: August 27, 2026 PT ยท Update: Superseded by 1.0.189.

  • Fixed a MiniMax H3 failure after a model was unloaded and reloaded, present in 1.0.181 and 1.0.182.
Docker images
  • CUDA 12 tag โ€” sha256:f51495b88cce27e6e7e3f7c625491202e6ca6e80a88039e822ab7f2790c7fc25
  • CUDA 13 tag โ€” sha256:26d64e40fb20b286ee487922e3b93cb4899696e4c905a23bc2f426ca99397e8e

#Older releases

Every version below is superseded. A worker updates straight to the current release; it never needs an intermediate version.

Version Released (PT) What it did
1.0.182 Aug 27, 2026 Attempted a GPU memory repair; the MiniMax H3 reload failure remained until 1.0.183.
1.0.181 Aug 27, 2026 Restored MiniMax H3 compatibility after the Balanced rollout.
1.0.180 Aug 27, 2026 Introduced MiniMax H3 Balanced; H3 Turbo moved to its LightX2V v1.1 files.
1.0.179 Aug 26, 2026 Faster MiniMax H3 reference-to-video; consistent high-quality video encode.
1.0.178 Aug 24, 2026 Managed-upgrade and MiniMax H3 compatibility maintenance.
1.0.177 Aug 23, 2026 First release eligible for MiniMax H3 reference-video jobs.
1.0.176 Aug 20, 2026 User LoRAs on every MiniMax H3 mode; better startup recovery.
1.0.175 Aug 19, 2026 MiniMax H3 Turbo moved to its native 768p acceleration file.
1.0.174 Aug 18, 2026 Restored the image and video jobs 1.0.173 broke; added the worker health data shown in the Dashboard.
1.0.173 Aug 18, 2026 Withdrawn: a hardening change blocked valid image and video jobs.
1.0.172 Aug 15, 2026 RTX VSR upscaling up to 16K.
1.0.171 Aug 14, 2026 First release eligible for RTX VSR and MiniMax H3 reference-to-video Turbo.
1.0.170 Aug 14, 2026 Introduced MiniMax Music 3, H3 reference-to-video Turbo and RTX VSR.
1.0.169 Aug 13, 2026 Introduced the LTX 2.5 family.
1.0.168 Aug 12, 2026 MiniMax H3 out-of-memory recovery leaves the GPU clean.
1.0.167 Aug 12, 2026 Automatic CUDA 13 image selection; less GPU memory spill with quantized LoRAs.
1.0.166 Aug 10, 2026 Rapid follow-up build; no operator action.
1.0.165 Aug 10, 2026 MiniMax H3 end-frame guidance and validated sampler choices.
1.0.164 Aug 8, 2026 MiniMax H3 Turbo default sampler restored.
1.0.163 Aug 7, 2026 First release eligible for the full MiniMax H3 Turbo set.
1.0.162 Aug 7, 2026 First H3 Turbo build; replaced within hours.
1.0.161 Aug 7, 2026 Short-lived rebuild; superseded the same day.
1.0.160 Aug 6, 2026 Krea 2 Identity Edit v0.3; MiniMax H3 keyframes fitted automatically.
1.0.159 Aug 5, 2026 MiniMax H3 export and reconnect reliability.
1.0.158 Aug 4, 2026 MiniMax H3 fixed at 24 fps; ACE-Step 1.5 XL on compatible 20 GB workers.
1.0.157 Aug 4, 2026 First release eligible for MiniMax H3 reference-to-video with audio.
1.0.156 Aug 4, 2026 MiniMax H3 execution and time-estimate refinements.
1.0.155 Aug 3, 2026 First correct native MiniMax H3 build.
1.0.154 Aug 3, 2026 Withdrawn within minutes. Do not install.
1.0.152 Aug 2, 2026 CUDA 12 only. Obsolete; do not install.
1.0.150 Aug 1, 2026 Wan Animate and large-LoRA memory fixes on RTX 4090 workers. CUDA 12 only.

1.0.151 and 1.0.153 have no public image.

Docker images for older releases

"Not published" means the release predated the public CUDA 13 track.

Version CUDA 12 image digest CUDA 13 image digest
1.0.182 sha256:115dd395b0eff22c7342a053607a88288ad79d6a2fc849a63f3c59c833d73b01 sha256:13004e43541f23507798da3786bf62ee75adcd43d2873cc27a057814d9f4b88c
1.0.181 sha256:e1216e06bc556640af73f67662e5260e53bcc4e99dbaa1e6723c5ab21e6d51a0 sha256:289014a81b03f314c6185ecaeb0f486cff7125ed9192368a86e442d5df511b70
1.0.180 sha256:a376b9d8dadba289aac963af02a9b312a0353c89b44008e6d259277112bc85cc sha256:1645ca7e68c86f0891908a6155831ca93f58235d4250079a5f358a5d13375ada
1.0.179 sha256:13f4f4dc5c2e319cc2aacb50bfc7f0ad73e3ae7ba0941233076669606703baad sha256:de80faea3a8423ddaa823320f1339eabf1e600c9babc402ca0e2acae7d6420e2
1.0.178 sha256:225b6e063cfb18eaf9f559307eb915c822541d5e1522426af69ae852e0b56201 sha256:43d177ca14967b18c723282a45d6a475a99150fa89ac03d8f7a54fff8714dff7
1.0.177 sha256:ba717bff20266fc111e979339f92ec48a10d76587b701670d468f2e2363650ff sha256:3bd9785a7fd492f9a2d19da2c7210334743f292331c47d98bbff8a41a4bc575a
1.0.176 sha256:36dac8d9a5c3bd7fd2a57c6ecf90888f09dc9de07e6563be8aa5b0c1d0bded91 sha256:6699c29457235b1d57f23b23184abc8f3ab4a145c5a58da10eb67a4cd2b6c95a
1.0.175 sha256:eab960826465f0b35e2fa4a75b713d2e5d999c58353412374111d9d855833daa sha256:f7d9d94093c2d03aaf9c79639ccff39d2681035f258b8a4bdb815536d5dcb7ed
1.0.174 sha256:30954a994dd7d287991fa03b551227b1c86fedea96b3ddc4947f231a3a2ebce1 sha256:63c2ca53b1dfa589cbb2240f5430b4a7ea3a2f08c4c035a2b0ef1ef722ddb1af
1.0.173 sha256:91e99482d5162316f8f59c378cfa42d204b95268302545f89707430f7f9cf836 sha256:56131949d460af38c6b9e3e491ecdeea88534fcc4290822a87774da1c992c741
1.0.172 sha256:db887ae99d4520f8d54926c7d93da630c3b855ac0bca7f07efe1955090a6d65c sha256:5174eccb9738170ecd3208443d23710ac8a5c7e164483ebb66d8d36e4b196738
1.0.171 sha256:216638a2c639f4bd5dbfb3c81ffbe1add0f506cc0dc25b4b2e9eef1e552f0099 sha256:1a3a6accee29539de312ebcaa9cc827ca3a3202d5b160056860375695b3483b8
1.0.170 sha256:12507a10f0ce61f0f53ae8e7b9c792e4c701e767a52309db46cfde75680badbc sha256:41a02f114e92121e03c84231fa7017c8d0192f9c3a77e8d0ec8f094814bc2b48
1.0.169 sha256:485f2f4cab39ddb83d35318c0105c3a11d1c1baf2b46c8b671b01ece9b883009 sha256:49eb8b406ab26b89c325eebeeb22e1a15153c80fec0579155200a9c66d4245b5
1.0.168 sha256:b14af703a9f1345a8d93e8462fbf514734a99557bb07ff17f65b2091bf58751b sha256:1304c6f0c5fb57bb4a8c845c8eef4023a2974b1643876663148a97c1402ce8b8
1.0.167 sha256:9ebc3b7bebd3d195d4f08845661b50c9599d79fc90d7e0364f42cb58da6fbc51 sha256:d36962a1ae2eed024e2967a6f8f2f463b86af733f4899ab9ffa1048e93fea127
1.0.166 sha256:461cc7986a41cc8fc9e755cca5401e763624e8c535d8b8b31bd3179208b2e682 sha256:3c997a03f08e9aed273b51866fa140e2360c5b3727329442012d7b3cd13db6ff
1.0.165 sha256:2f0a4cc5022cdbb9f9d85233d2fef88ce173c2e527df14a740ad7bc7c4c48b8f Not published
1.0.164 sha256:6526b2f1b1d4782b2125c0642ee9291a533b2f81ae1b284014491ab62a73823c Not published
1.0.163 sha256:07b5a48404c7895a21a12f45ca3f77cfabf09d9bc444a5491a8cb8e9ec9ea539 Not published
1.0.162 sha256:bb72fe5f0e9243d625043338646e6b2dc2550683d32492d4ba9cdc208a9a5e51 Not published
1.0.161 sha256:1575fd19b412a6ebdb6899aa295705e77cb51dd489949199fd8d72d16b74ef1f Not published
1.0.160 sha256:86ca11610410654976ea3a998754da8c151585bee86c2c4e91b869054e37f37e Not published
1.0.159 sha256:6e26b20e01ebe1cc7802a711c7a4b140ee2f1b52b518554bfc4337844ee9a371 Not published
1.0.158 sha256:31f5a6f9dc1bd516270e8b9e1aa04576a04a89b1d7f07bcb2f5af6cb20eb90f5 Not published
1.0.157 sha256:0d15ed9a8f4fc290093a9af15ced3fe08b3422d06bd5acd9ad7755bb119337a3 Not published
1.0.156 sha256:226990ffacbab87c5a9cca12e9483ddde7848ec043fe4701322e6287343fd0ff Not published
1.0.155 sha256:8daeebc88f16cbdf010691388ef39e8b427b149272ee562f2c50c5f17331ef53 Not published
1.0.154 sha256:40fa18efc12e28ec2955a117b7dfd2ecfb8d3d78c94959fb6d5df34ef35839ba Not published
1.0.152 sha256:7292e42a8d98a8b47464e4179f29ca8f615f36f2fa9d52324910194884d2d7b1 Not published
1.0.150 sha256:867b8aaae5cdc8fb23423bba555b3f236be98d946b35fe07bce5a8ac63c986ce Not published

Full tag histories: CUDA 12 and CUDA 13.

Last updated 2026-11-20