You can deploy this model and start earning money today!

mlx-community/
Tmax-9B-MLX-bf16

Name: Tmax-9B-MLX-bf16
Availability: InStock

Views

No views yet

Tmax-9B MLX (bf16)

MLX-converted text-only weights of allenai/tmax-9b.

The upstream base ships as a multimodal Qwen3_5ForConditionalGeneration config but contains zero vision tensors in its safetensors — i.e. it is already a text-only checkpoint with stub vision metadata. This release strips the residual vision_config / image-token entries so it loads cleanly via mlx_lm without a vision tower.

Source: allenai/tmax-9b
License: Apache-2.0
Variant: bf16
Quantized by: raullenchai
Tooling: mlx-lm 0.31.3 (the upstream mlx_vlm 0.3.12 qwen3_5 loader hard-requires vision-tower weights that this base does not ship, so the text-only mlx_lm.convert path is used instead)
Chat template: ships with the source repo (chat_template.jinja)
Tool format: qwen3_xml-compatible (<tool_call>{json}</tool_call>)

Usage

python

from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Tmax-9B-MLX-bf16")
print(generate(model, tokenizer, prompt="Hello", max_tokens=32))

Notes

This is a pure text-generation MLX release. No vision/image inputs.
For best chat behavior, use the chat template that ships with this repo.

Benchmarks

Measured on M3 Ultra Studio (28 (20 Performance and 8 Efficiency) CPU, 60-core GPU, 256 GB unified memory) via rapid-mlx 0.8.18. Medians of 3 runs.

Variant	Decode tok/s	TTFT (ms)	Prefill 1k (tok/s)	Prefill 4k (tok/s)	Prefill 16k (tok/s)	Tool-call e2e
Tmax-9B (bf16 MLX)	—	—	—	—	—	—

⚠️ Note: weights load successfully, but the first streaming TTFT call never returned in our bench harness (43 min, watchdog killed). This is bf16-specific — the 4 / 6 / 8 bit Tmax-9B variants stream cleanly. Recommend the quantized variants for streaming workloads until root-caused.

Full results (all 7 Tmax MLX variants + 2 Qwen3.5 controls): rapid-mlx docs.

Reproduce:

bash

pip install rapid-mlx==0.8.18
rapid-mlx serve tmax-9b-bf16 --port 8765