Search results

Failed to run deepseek-ocr model by api(/v1/chat/completion)

Status: Open.

#244 In lmstudio-ai/mlx-engine;

· zutuanwang opened on Nov 21, 2025

Fail to load model qwen3-omni

Status: Open.

#243 In lmstudio-ai/mlx-engine;

· zutuanwang opened on Nov 21, 2025

[qwen3 vl] [qwen2 vl] Add multimodal RoPE and Deepstack implementation to mlx-lm language model port

bugSomething isn't workingSomething isn't working

Status: Open.

#237 In lmstudio-ai/mlx-engine;

· neilmehta24 opened on Oct 23, 2025

Image tokens not always included when running Qwen3-VL

bugSomething isn't workingSomething isn't working

fixed-in-next-releaseThe next release of LM Studio fixes this issueThe next release of LM Studio fixes this issue

Status: Open.

#233 In lmstudio-ai/mlx-engine;

· vincentamato opened on Oct 14, 2025

Airllm support techniques

Status: Open.

#229 In lmstudio-ai/mlx-engine;

· Isaina opened on Oct 4, 2025

DEBT: Prefill Trickery

Status: Open.

#226 In lmstudio-ai/mlx-engine;

· will-lms opened on Sep 25, 2025

[Docs/Question] How to build and use custom mlx-engine with updated mlx-lm for Qwen3-Next-80B-A3B-Instruct-8bit?

Status: Open.

#220 In lmstudio-ai/mlx-engine;

· do-me opened on Sep 15, 2025

[Model Request] Add support for Kimi-K2 model family

Status: Open.

#219 In lmstudio-ai/mlx-engine;

· KarlXMagnusson opened on Sep 10, 2025

Removing external code when models are upstreamed into Transformers

Status: Open.

#211 In lmstudio-ai/mlx-engine;

· christian-lms opened on Aug 19, 2025

[ERROR] Error: Channel Error when submitting long input (>256KiB for llama 4) for models with large context (e.g. llama 4 with >1M) on Mac Studio M3 Ultra (512GiB RAM)

Status: Open.

#208 In lmstudio-ai/mlx-engine;

· KarlXMagnusson opened on Aug 17, 2025

Cross-Post: Images processed with MLX vision models are downsized to 500px long edge

Status: Open.

#206 In lmstudio-ai/mlx-engine;

· bmachek opened on Aug 10, 2025

Status of MTP for GLM?

enhancementNew feature or requestNew feature or request

Status: Open.

#205 In lmstudio-ai/mlx-engine;

· PhillipVas opened on Aug 6, 2025