Opper TaskBench — leaderboard
Metric: Avg Task Score (%). Source: opper.ai. 86 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5 | 96.4 |
| 2 | Claude Opus 4.7 | 95.4 |
| 3 | Gemini 3.1 Pro (Preview) | 95.3 |
| 4 | Gemini 3 Flash (Preview) | 95.1 |
| 5 | Gemini 3 Pro (Preview) | 94.8 |
| 6 | Gemini 3.5 Flash | 94.7 |
| 7 | Claude Opus 4.8 | 94.5 |
| 8 | GPT-5.3 Codex | 94.5 |
| 9 | Claude Opus 4.6 | 94.4 |
| 10 | Qwen 3.7 Max | 93.3 |
| 11 | DeepSeek V4 Pro | 93.1 |
| 12 | Claude Opus 4.5 | 93 |
| 13 | GLM-5.2 | 93 |
| 14 | Kimi K2.5 | 92.2 |
| 15 | GLM-5 | 92.2 |
Interactive version: theaggregate.ai/benchmark?slug=opper-taskbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.