Views
No views yet
Qwen/Qwen-AgentWorld-35B-A3B at upstream revision 60d2b0434a53d2e62a7c00a489586815d94ebffb, converted from the BF16 safetensors with mlx_lm convert (mlx-lm 0.31.3). ~17.2 GiB on disk.Runtime status (verified 2026-07-04): loads and generates under mlx-lm 0.31.3 — chat and code probes passed on this exact payload. Republished 2026-07-04 in the mlx-lm layout, replacing the 2026-07-02 pack, which no MLX runtime could load. Output quality beyond the smoke probes is not verified.
| Runtime | Loads? | Notes |
|---|---|---|
| mlx-lm 0.31.3 | ✅ | chat + code smoke passed per variant |
| mlx-vlm 0.6.3 | n/a | text-only model (no vision tower) |
pipelines/vlm_pack_smoke.py / pipelines/lm_pack_smoke.py, majek repo); verdicts in .sisyphus/evidence/lane-a-republish/ and in PROVENANCE.md.1from mlx_lm import load, generate
2
3model, tokenizer = load("majentik/Qwen-AgentWorld-35B-A3B-MLX-MXFP4")
4prompt = tokenizer.apply_chat_template(
5 [{"role": "user", "content": "Explain KV caching in one sentence."}],
6 add_generation_prompt=True, tokenize=False,
7)
8print(generate(model, tokenizer, prompt=prompt, max_tokens=128))1# base = snapshot of Qwen/Qwen-AgentWorld-35B-A3B @ 60d2b0434a53d2e62a7c00a489586815d94ebffb
2python -m mlx_lm convert --hf-path /tmp/lane-a-republish/qwen-agentworld-35b-a3b/base --mlx-path /tmp/lane-a-republish/qwen-agentworld-35b-a3b/MXFP4 -q --q-bits 4 --q-group-size 32 --q-mode mxfp4majentik/Qwen-AgentWorld-35B-A3B-MLX-MXFP4 (this repo)Qwen/Qwen-AgentWorld-35B-A3B @ 60d2b0434a53d2e62a7c00a489586815d94ebffbpipelines.mlx_direct_quantize pack. Full details in PROVENANCE.md in this repo.Qwen/Qwen-AgentWorld-35B-A3B. All rights in the original model remain with its authors.