E-Bench - Pass^3: leaderboard
E-Bench (Tencent Hunyuan, 2026) builds synthetic product environments modelled on Honor of Kings, QQ Music and Tencent Meeting with 323 state-changing tasks that require discovering hidden data and composing several tool calls, graded by exact database-state diff over three trials per task. This board is Pass^3, the share of tasks solved in all three trials, in percent, with only the domain tools; higher is better.
Metric: Pass^3 (%; tasks solved in all three trials; domain MCP tools only) on 323 synthetic state-changing tasks in three product environments (Honor of Kings, QQ Music, Tencent Meeting), graded by exact database-state diff; three trials per task, each model at its highest available thinking effort in a shared MCP harness. Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K3 (Max) | 58.82 | #54 (Kimi K3) |
| 2 | GPT-5.5 (xHigh) | 57.59 | #49 (GPT-5.5) |
| 3 | Grok 4.5 (High) | 52.32 | #82 (Grok 4.5) |
| 4 | Claude Opus 4.8 (Max) | 50.81 | #56 (Claude Opus 4.8) |
| 5 | GLM-5.2 (Max) | 30.96 | #138 (GLM-5.2) |
| 6 | Qwen 3.7 Max | 30.34 | #145 |
| 7 | Seed 2.1 Pro | 28.79 | #135 |
| 8 | Hy3 | 26.63 | #195 |
| 9 | Gemini 3.5 Flash (High) | 21.98 | #101 (Gemini 3.5 Flash) |
| 10 | MiniMax-M3 | 20.12 | #269 |
| 11 | DeepSeek V4 Pro (Max) | 17.34 | #188 (DeepSeek V4 Pro) |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=e-bench-pass-pow-3 · How It Works · Data refreshed daily, snapshot 2026-11-13.