GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
axteam
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
axteam
Model Tree
Qwen/Qwen3.5-9B
Input Modalities
Output Modalities
Supported Functionality
pack-quantized.model.visual.*), the multi-token-prediction head (mtp.*), all RMSNorms, token embeddings, the LM head, the SSM non-linear parts (linear_attn.conv1d/A_log/dt_bias). Left unquantized to preserve quality (a small fraction of parameters).chat_template / generation_config are identical to the original. config.json is the original plus a quantization_config block.This is post-training quantization (PTQ) — no retraining, so behaviour stays faithful to the original. AWQ uses a small calibration set (no gradient) to protect salient weight channels before rounding.
Fidelity to the BF16 original on an internal mixed English/Korean set (100 prompts, BF16 greedy answers scored under each build — same harness for every row).
| Metric | This model | BF16 original |
|---|---|---|
| Perplexity (lower = closer) | 1.5403 | 1.4506 (+6.2%) |
| Top-1 next-token agreement | 90.9% | 99.2% |
These measure fidelity to the base model, not absolute task scores.
bash
vllm serve axteam/Qwen3.5-9B-AWQ-4bitRecommended sampling (Qwen):
temperature=0.7, top_p=0.8, top_k=20.
Quantized derivative of Qwen/Qwen3.5-9B (Apache-2.0). All original weights, tokenizer and chat template belong to the Qwen team.
Qwen/Qwen3.5-9B 를 AWQ INT4 (W4A16) 로 양자화한 서빙용(4-bit) 모델입니다. 작성: axteam.
pack-quantized.vllm serve axteam/Qwen3.5-9B-AWQ-4bit, 권장 샘플링 temperature=0.7, top_p=0.8, top_k=20.