Qwen3-30B-A3B-Instruct-2507-AWQ
AWQ quantization: done by stelterlab in INT4 GEMM with llm-compressor (https://github.com/vllm-project/llm-compressor) from the vllm-project. Still trying to get familiar with llm-compressor. Tested against vLLM v0.9.2.
Params
30.8 B
Context
262,144
Downloads 30d
411 K
Likes
1
Download history
daily snapshots · 27 days492 K191 K
Sep 28Oct 7Oct 16Oct 24
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| model.safetensors | i32 | 16.7 GB | 23.5 GB | ⚠️ Tight — reduce context |
| model.safetensors (bf16, full) | bf16 + 256K ctx | 16.7 GB | 166.5 GB | ❌ Won’t fit |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ curl -s https://aimodelscomparison.com/api/v1/models/qwen3-30b-a3b-instruct-2507-awq
{
"hf_id": "stelterlab/Qwen3-30B-A3B-Instruct-2507-AWQ",
"params_b": 30.77,
"context_length": 262144,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 410570,
"vram_estimates": [
{ "quant": "i32", "gb": 23.5 }
],
"updated_at": "2026-09-28T01:00:26Z"
}
Specifications
- Architecture
- Qwen3MoeForCausalLM
- Parameters
- 30.8 B
- Tensor type
- I32
- Context length
- 262,144
- Vocabulary
- 151,936
- Layers / heads
- 48 / 32
- Licence
- apache-2.0
- First seen on the Hub
- 2025-08-02
- Base model
- Qwen3-30B-A3B-Instruct-2507
- Training datasets
- undisclosed
- Added to our catalog
- 2026-09-28
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model