LongWebBench: leaderboard
Metric: Overall (self-reported). Source: benchmarklist.com. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 7.22 |
| 2 | Claude Sonnet 4.5 (Thinking) | 6.74 |
| 3 | Gemini 3 Flash (Preview) | 6.53 |
| 4 | GPT-5.2 | 6.5 |
| 5 | Claude Opus 4.5 (20251101) | 6.43 |
| 6 | GLM-4.6V | 5.96 |
| 7 | Qwen 3 VL 235B A22B Instruct | 5.85 |
| 8 | Doubao-Seed-1.6 | 4.77 |
| 9 | Qwen 3 VL 8B Instruct | 3.95 |
| 10 | InternVL3-78B | 3.62 |
| 11 | GPT-4o | 3.53 |
Interactive version: theaggregate.ai/benchmark?slug=longwebbench · How It Works · Data refreshed daily, snapshot 2026-09-05.