Instructions to use mlx-community/Tmax-27B-MLX-6bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlx-community/Tmax-27B-MLX-6bit with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Tmax-27B-MLX-6bit")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use mlx-community/Tmax-27B-MLX-6bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/Tmax-27B-MLX-6bit"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "mlx-community/Tmax-27B-MLX-6bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use mlx-community/Tmax-27B-MLX-6bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "mlx-community/Tmax-27B-MLX-6bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "mlx-community/Tmax-27B-MLX-6bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "mlx-community/Tmax-27B-MLX-6bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use mlx-community/Tmax-27B-MLX-6bit with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/Tmax-27B-MLX-6bit"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default mlx-community/Tmax-27B-MLX-6bit

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use mlx-community/Tmax-27B-MLX-6bit with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/Tmax-27B-MLX-6bit"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "mlx-community/Tmax-27B-MLX-6bit" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Raullen commited on Jun 25

Commit

2a8d8f4

verified ·

1 Parent(s): 93d8d16

bench: correct 27B prefill_16k numbers (false hang ruled out, hybrid attention narrative)

Browse files

Files changed (1) hide show

README.md +3 -3

README.md CHANGED Viewed

@@ -45,13 +45,13 @@ print(generate(model, tokenizer, prompt="Hello", max_tokens=32))
 ## 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.
 | Variant | Decode tok/s | TTFT (ms) | Prefill 1k (tok/s) | Prefill 4k (tok/s) | Prefill 16k (tok/s) | Tool-call e2e |
 |---|---:|---:|---:|---:|---:|---:|
-| Tmax-27B (6-bit MLX) | 26.8 | 288 | 305 | 314 | — | — |
-> ⚠️ **Note**: 16k-context streaming currently hangs in our bench harness on all Tmax-27B quantizations. Shorter contexts (≤4k) work fine. Investigation pending — see [issue](https://github.com/raullenchai/Rapid-MLX/issues) for status.
 Full results (all 7 Tmax MLX variants + 2 Qwen3.5 controls): [rapid-mlx docs](https://github.com/raullenchai/Rapid-MLX/blob/main/docs/benchmarks/tmax-m3-ultra.md).

 ## 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-27B (6-bit MLX) | 26.8 | 288 | 305 | 314 | 303 | 2489 ms (OK) |
+> **Architecture note**: Tmax-27B uses a hybrid Gated-DeltaNet design (3:1 linear-attention to full-attention layer mix). 16k-context prefill is bandwidth-bound at ~310 tok/s regardless of quantization bit width — ~54 s wall to first token at 16k. This is an architectural property of hybrid linear-attention models on Apple Silicon, not a regression, and not a rapid-mlx bug. Decode and short-context (≤4k) tool-call performance are competitive with the dense Qwen3.5-27B-4bit control on the same hardware.
 Full results (all 7 Tmax MLX variants + 2 Qwen3.5 controls): [rapid-mlx docs](https://github.com/raullenchai/Rapid-MLX/blob/main/docs/benchmarks/tmax-m3-ultra.md).