Qwen3-235B-A22B-Instruct-2507-FP8
We introduce the updated version of the Qwen3-235B-A22B-FP8 non-thinking mode, named Qwen3-235B-A22B-Instruct-2507-FP8, featuring the following key enhancements:
Params
235.1 B
Context
262,144
Downloads 30d
382 K
Likes
150
Download history
daily snapshots · 17 days382 K188 K
Sep 4Sep 9Sep 15Sep 20
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 | f8_e4m3 | 236.4 GB | 295.8 GB | ❌ Won’t fit |
| model.safetensors (bf16, full) | bf16 + 256K ctx | 236.4 GB | 1,389.1 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/qwen3-235b-a22b-instruct-2507-fp8
{
"hf_id": "Qwen/Qwen3-235B-A22B-Instruct-2507-FP8",
"params_b": 235.11,
"context_length": 262144,
"license": { "id": "apache-2.0", "commercial": "yes" },
"downloads_30d": 382444,
"vram_estimates": [
{ "quant": "f8_e4m3", "gb": 295.8 }
],
"updated_at": "2026-09-04T01:00:25Z"
}
Specifications
- Architecture
- Qwen3MoeForCausalLM
- Parameters
- 235.1 B
- Tensor type
- F8_E4M3
- Context length
- 262,144
- Vocabulary
- 151,936
- Layers / heads
- 94 / 64
- Licence
- apache-2.0
- First seen on the Hub
- 2025-07-21
- Training datasets
- undisclosed
- Added to our catalog
- 2026-09-04
Compare with any text-generation model
Popular comparisons
Appears in