Qwen3 30B-A3B Instruct 2507 scores 11.5 on the Respan Index (#134), ahead of Llama 3.3 70B Instruct at 8.6 (#137). Qwen3 30B-A3B Instruct 2507 leads on 4 of the 4 benchmarks both report. Qwen3 30B-A3B Instruct 2507 has the larger context window (262K tokens).
| Benchmark | Llama 3.3 70B Instruct | Qwen3 30B-A3B Instruct 2507 |
|---|---|---|
| GPQA Diamond | 50.5% | 55.6% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 5.1% | 62.2% |
| LMArena Elo | 1274 | 1383 |
| MMLU-Pro | 68.9% | 78.4% |
Higher is better. Each score links to where it was published.
* Estimated: no published scores in that category. How the Respan Index works