All issues

Search results

MLX server (#353) as an Engine Protocol runtime: use case + offer to help

Status: Open.

#373 In lmstudio-ai/mlx-engine;

· archmagos-certiv opened on Sep 2, 2026

glm5_next (GLM-5.3-Flash): prompt cache save always fails — save_safetensors on zero-size array in batched-vision blob store; plus latent hybrid-cache bug in CacheWrapper._num_tokens_in_cache

Status: Open.

#372 In lmstudio-ai/mlx-engine;

· tobieapb opened on Aug 31, 2026

Draft-only (MTP) repos indexed as loadable models, then crash at generation: 'Qwen3_5MTPDraftModel' object has no attribute 'get_input_embeddings'

Status: Open.

#371 In lmstudio-ai/mlx-engine;

· kafasam opened on Aug 30, 2026

[Feature] MLX engine ignores built-in MTP (nextn) heads — native MTP speculative decoding missing on the MLX path

Status: Open.

#370 In lmstudio-ai/mlx-engine;

· nikolai-vysotskyi opened on Aug 21, 2026

Batched vision models cannot use a draft model: is_draft_model_compatible() returns False unconditionally

Status: Open.

#368 In lmstudio-ai/mlx-engine;

· cristyanul opened on Aug 19, 2026

MLX auto-fit replaces the configured context length and disagrees with the prompt cache budget

Status: Open.

#366 In lmstudio-ai/mlx-engine;

· philippdubach opened on Aug 15, 2026

Partial GPU Offloading support for MLX Qwen3.6-35b-a3b 8bit on M4 Mac 48GB?

Status: Open.

#365 In lmstudio-ai/mlx-engine;

· combacsa opened on Aug 13, 2026

Backend scheduler crashes with broadcast_shapes ValueError in BatchRotatingKVCache.merge under concurrent requests with shared prefix

Status: Open.

#363 In lmstudio-ai/mlx-engine;

· seb760 opened on Aug 7, 2026

repeat_penalty causes fatal backend-scheduler crashes in multi-turn conversations (32-bit slice error / KV cache merge broadcast error)

Status: Open.

#359 In lmstudio-ai/mlx-engine;

· thefrenchbaron opened on Aug 5, 2026

Text-only architectures implemented only in mlx_vlm (e.g. deepseek_v4) fail to load because dispatch checks only for vision_config

Status: Open.

#357 In lmstudio-ai/mlx-engine;

· True2456 opened on Aug 1, 2026

Disk prefix KV caching

Status: Open.

#354 In lmstudio-ai/mlx-engine;

· richardwerkman opened on Jul 24, 2026

Please support for Ternary Bonsai 27b with vision an reasoning

Status: Open.

#347 In lmstudio-ai/mlx-engine;

· lustfeind opened on Jul 15, 2026