Qwen3.6 27B scores 20.8 on the Respan Index (#91), ahead of GPT-5.4 nano at 20.5 (#93). GPT-5.4 nano leads on 7 of the 13 benchmarks both report. GPT-5.4 nano has the larger context window (400K tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 70.8% |
| 71.8% |
| LiveBench Data Analysis (LiveBench) | 67.6% | 70.4% |
| LiveBench Instruction Following (LiveBench) | 67.2% | 53.2% |
| LiveBench Language (LiveBench) | 62.5% | 63.3% |
| LiveBench Mathematics (LiveBench) | 91% | 79.9% |
| LiveBench Reasoning (LiveBench) | 81.1% | 70.3% |
| Mystery Game Puzzles (Epoch AI) | 9% | 7% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 87.8% | 91.1% |
| SAGE (Vals AI) | 38.1% | 45.6% |
Higher is better. Each score links to where it was published.