OsaurusAI/Ornith-1.5-35B-A3B-MXFP8
Ornith-1.5-35B-A3B-MXFP8 MLX bundle of ornith-ai/Ornith-1.5-35B-A3B — .
Ornith 1.5 is an agentic coding / reasoning VLM built on a hybrid gated-delta linear attention + full attention backbone (3:1), with a 27-layer vision tower and native video support.
Bundle
| Field | Value |
|---|---|
| Source | ornith-ai/Ornith-1.5-35B-A3B |
| Architecture | qwen3_5_moe / Qwen3_5MoeForConditionalGeneration |
| Size on disk | 34.95 GiB |
| Layers | 40 |
| Hidden size | 2048 |
| Routed experts | 256 |
| Context | 262,144 |
| Shards | 33 |
| Quantization | uniform 8-bit mxfp8, group size 32 |
How it was quantized
Three calibration methods, all driven by one capture pass — the per-input-channel second moment E[x_c^2] is simultaneously the Hessian diagonal, the imatrix weighting and the AWQ salient-channel statistic.
| Method | What it does here |
|---|---|
| Hessian-trace allocation | Bits assigned by measured tr(H)·‖W‖²_F per module, not by tensor name. The vision tower scores higher than the text MLP on this model, which a name-based profile gets backwards. |
| imatrix refit | Activation-weighted affine fit replacing RTN codes. |
Tensors whose in_features is divisible by no MLX group size (the 27 vision linear_fc2 at 4304) stay fp16.
Modalities
| Modality | Status |
|---|---|
| Text | supported |
| Vision | supported — 333 vision-tower tensors, preprocessor_config.json + processor_config.json ride with the bundle |
| Video | supported — video_preprocessor_config.json present; verified end-to-end |
| Audio | not supported. The tokenizer defines `< |
Reasoning
Reasoning is ON by default — the no-kwarg generation prompt is byte-identical to enable_thinking=True and ends <|im_start|>assistant\n<think>\n.
It is toggleable, but note how: enable_thinking=False does not remove the think block, it prefills an empty closed one (<think>\n\n</think>\n\n). A parser testing merely for the presence of a <think> block will find one in both modes — test whether it has content.
There are no reasoning_effort tiers on this model family (unlike Qwen3.8). History <think> blocks are preserved unconditionally. Reasoning parser: qwen3; tool parser: qwen3_coder.
Sampling
Both presets from the vendor card are stamped into jang_config.json, and the coding preset is also written to generation_config.json so the two files agree.
Ornith 1.5 is an agentic coding model (SWE-bench Verified 79, Terminal-Bench 2.1 67.8), so this bundle defaults to the coding preset. Upstream's own
generation_config.jsonships the general numbers (temp 1.0, presence 1.5) — usesampling_modes.generalif you want parity with the vLLM/Transformers defaults.
| Preset | temp | top_p | top_k | min_p | presence | repetition |
|---|---|---|---|---|---|---|
| general | 1.0 | 0.95 | 20 | 0.0 | 1.5 | 1.0 |
| coding (default) | 0.6 | 0.95 | 20 | 0.0 | 0.0 | 1.0 |
Stop tokens: [248046, 248044] (<|im_end|>, <|endoftext|>).
Speculative decoding (MTP)
This bundle preserves the native MTP head (1561 mtp.* tensors). Recommended 1 draft/step on Apple silicon (vmlx_mtp_tuning.json); that is a recommendation, not a measured sweep on this artifact.
Credits
JANG quantization by Jinho Jang — eric@osaurus.ai
Base model: ornith-ai/Ornith-1.5-35B-A3B by Ornith AI.