Qwen3-235B-A22B-Thinking (Jul 2025)
Qwen3-235B-A22B-Thinking (Jul 2025) is an open-weights language model from Qwen, released July 25, 2025. It has 235B parameters (mixture-of-experts, 22B active). On the Epoch Capabilities Index it scores 143.2, rank 107 of 256 rated models.
Today, we are excited to announce the release of Qwen3, the latest addition to the Qwen family of large language models. Our flagship model, Qwen3-235B-A22B, achieves competitive results in benchmark evaluations of coding, math, general capabilities, etc., when compared to other top-tier models such as DeepSeek-R1, …
Check it against your own PC →
Can you run Qwen3-235B-A22B-Thinking (Jul 2025) locally?
The smallest common GPU it runs on is 32 GB of VRAM (with 32 GB of RAM): IQ1_M, GPU + experts in RAM, roughly 20 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Too large | – | – | – | – |
| 8 GB VRAM | Too large | – | – | – | – |
| 12 GB VRAM | Too large | – | – | – | – |
| 16 GB VRAM | Too large | – | – | – | – |
| 24 GB VRAM | Too large | – | – | – | – |
| 32 GB VRAM | Runs · GPU + CPU experts | IQ1_M (experimental) | 31.2 GB | 24.8 GB | ~20 |
Strengths by category
ModelTape's 0–100 fit on Epoch AI benchmark results.
| Category | Score | Based on |
|---|---|---|
| Math | 37.8 | OTIS Mock AIME 2024-2025, FrontierMath-2025-02-28-Private, FrontierMath-Tier-4-2025-07-01-Private |
| Knowledge | 67.1 | GPQA diamond, SimpleQA Verified |
Benchmark results
| Benchmark (Epoch AI) | Score |
|---|---|
| OTIS Mock AIME | 86.7% |
| GPQA Diamond | 76.5% |
| SimpleQA Verified | 40.4% |
| FrontierMath-2025-02-28-Private | 14.9% |
| FrontierMath-Tier-4-2025-07-01-Private | 0.0% |
| Mystery Game Puzzles | 0.0% |
Links
Announcement · Epoch AI benchmarks
Models with a similar score
o3 · Nemotron 3 Ultra · Grok 4 · DeepSeek-V4-Flash · Qwen3.6 27B · GLM-5 · GPT-5.4 Nano · DeepSeek-V3.2
More from Qwen
Qwen3.8 Omni Flash · Qwen3.8 Max (0902) · Qwen3.8 Flash · Qwen3.8 Flash Next · Qwen 3.8 27B · Qwen3.8-2.4T-A95B · Qwen 3.8 Max · Qwen3.7 Flash