Qwen3.8 27B scores 40.5 on the Respan Index (#45), ahead of GLM-5.3-Flash at 40.4 (#47). GLM-5.3-Flash leads on 9 of the 14 benchmarks both report. GLM-5.3-Flash is 4.7x cheaper per token at list price. GLM-5.3-Flash has the larger context window (1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 76.4% |
| 76.6% |
| LiveBench Instruction Following (LiveBench) | 52.8% | 72.7% |
| LiveBench Language (LiveBench) | 77.3% | 74.3% |
| LiveBench Mathematics (LiveBench) | 81.2% | 86.2% |
| LiveBench Reasoning (LiveBench) | 77.6% | 80% |
| LMArena Agent (LMArena) | -0.0044 | -0.0166 |
| LMArena Elo | 1472 | 1439 |
| LMArena Vision (LMArena) | 1278 | 1243 |
| LMArena WebDev (LMArena) | 1616 | 1591 |
| Terminal-Bench 2.1 | 84.3% | 73% |
Higher is better. Each score links to where it was published.