FORTE - Pass^3: leaderboard
FORTE office-productivity agent benchmark (AGI-Eval): Pass^3 is the share of tasks solved in all three independent OpenClaw runs, a consistency measure under all-or-nothing LLM-judge rubric grading with thinking enabled at the highest available effort. %, higher is better.
Metric: Pass^3 (%, task solved in all three runs). Source: agi-eval-official.github.io. Saturation forecast: Around February 2027. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 69.95 | #39 |
| 2 | GPT-5.5 | 68.85 | #33 |
| 3 | Claude Opus 4.7 | 68.85 | #51 |
| 4 | GLM-5.2 | 63.93 | #80 |
| 5 | Claude Opus 4.6 | 63.39 | #67 |
| 6 | GLM-5.1 | 61.2 | #109 |
| 7 | Gemini 3.1 Pro (Preview) | 60.11 | #64 |
| 8 | LongCat 2.0 | 59.02 | #220 |
| 9 | Claude Sonnet 4.6 | 56.83 | #94 |
| 10 | GPT-5.4 | 56.83 | #83 |
| 11 | DeepSeek V4 Flash | 56.28 | #155 |
| 12 | MiniMax-M3 | 56.28 | #151 |
| 13 | DeepSeek V4 Pro | 54.64 | #112 |
| 14 | Kimi K2.6 | 46.45 | #113 |
| 15 | Qwen 3.6 Plus | 45.9 | #124 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=forte-pass-pow-3 · How It Works · Data refreshed daily, snapshot 2026-11-05.