AGI-Eval Community - Objective Accuracy: leaderboard
AGI-Eval community leaderboard (agi-eval.cn), the private-set LLM board built from the platform's non-public items; latest monthly snapshot. Unrelated to the AGIEval exam dataset. This board is overall objective-evaluation accuracy across all ability dimensions (the board's ranking column). Upstream accuracy 0-1, shown as 0-100; higher is better.
Metric: Accuracy (%). Source: agi-eval.cn. Saturation forecast: Around February 2027. 141 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 86.14 | #33 |
| 2 | Claude Fable 5 | 85.21 | #20 |
| 3 | Gemini 3.1 Pro (Preview) | 84.05 | #64 |
| 4 | Kimi K3 | 83.73 | #35 |
| 5 | Gemini 3.5 Flash | 82.82 | #66 |
| 6 | Seed 2.1 Pro | 82.56 | #73 |
| 7 | Claude Opus 4.8 | 81.78 | #39 |
| 8 | GPT-5.4 | 81.62 | #83 |
| 9 | Qwen 3.7 Max | 81.61 | #85 |
| 10 | Seed 2.0 Pro | 81.42 | #122 |
| 11 | DeepSeek V4 Pro (Max) | 81.3 | #112 (DeepSeek V4 Pro) |
| 12 | Kimi K2.6 | 81.21 | #113 |
| 13 | Claude Opus 4.6 | 80.71 | #67 |
| 14 | Claude Sonnet 5 | 80.42 | #65 |
| 15 | GLM-5.2 | 80.41 | #80 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=agi-eval-community-objective-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-05.