Up to four models
GPT-5 mini vs Qwen3-235B-A22B-Thinking (Jul 2025)
On the index (GPT-4 = 100), GPT-5 mini scores 140 and Qwen3-235B-A22B-Thinking (Jul 2025) 138. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
GPT-5 mini
140124–156
WeightsClosed
ReleasedAug 2025
Price in / out$0.25 / $2.0
Context400K
Core tests taken5 of 9
138121–154
WeightsOpen
ReleasedJul 2025
Price in / out$0.45 / $3.5
Context128K
Core tests taken1 of 9
GPT-5 mini and Qwen3-235B-A22B-Thinking (Jul 2025) are too close to call (2 points apart, within the uncertainty), ahead on 0 of 1 shared core tests, and costs 1.8× less.
Test by testthird-party results
GPT-5 mini
Qwen3-235B-A22B-Thinking (Jul 2025)
Every result as a table
| Test | GPT-5 mini | Qwen3-235B-A22B-Thinking (Jul 2025) |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 75.0% | 80.0% |
| Fiction.LiveBench | 69.4% | 75.0% |
| SimpleQA Verified | 21.6% | 40.4% |
| Maths | ||
| AIME-style maths (OTIS mock) | 86.7% | 86.7% |
| FrontierMath (2025 set) | 27.2% | 8.5% |
| FrontierMath tier 4 (2025 set) | 6.3% | 0.0% |
| Coding | ||
| WeirdML | 52.7% | 41.0% |
| Agents | ||
| DTBench | 80.5% | 80.3% |
| LMCA | 34.2% | 29.3% |
| Mystery game puzzles | 10.0% | 9.0% |
Tests at least two of these models have taken. Bold rows are the core tests.