QwenWebBench — Qwen3.6-card web-agent bench, non-% scale (median 1197, Elo-like); distinct from QwenClawBench.
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.5 | Anthropic | 1536 | 1 | 2026-06-15 |
| 2 | Qwen3.6 27B | Alibaba | 1487 | 1 | 2026-06-15 |
| 3 | Qwen3.6 35B A3B | Alibaba | 1397 | 2 | 2026-06-15 |
| 4 | Gemma 4 31B | 1197 | 2 | 2026-06-15 | |
| 5 | Qwen3.5 397B A17B | Alibaba | 1186 | 1 | 2026-06-15 |
| 6 | Gemma 4 26B A4B | 1178 | 1 | 2026-06-06 | |
| 7 | Qwen3.5 27B | Alibaba | 1068 | 2 | 2026-06-15 |
| 8 | Qwen3.5 35B A3B | Alibaba | 978 | 1 | 2026-06-06 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.