GameHorizon - Offline T2 Multi-Horizon Decomposition: leaderboard
Metric: Accuracy (%). Source: gamehorizon-suite.github.io. Saturation forecast: Around February 2027. 43 models tracked. Updated 18 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.8 Flash | 81.2 | #39 |
| 2 | Gemini 3.7 Flash | 80.1 | #38 |
| 3 | GPT-6 Astra | 79.6 | #7 |
| 4 | Seed 2.1 Turbo | 79.4 | #165 |
| 5 | Kimi K3 | 79.3 | #60 |
| 6 | Gemini 3.6 Flash | 79.1 | #77 |
| 7 | GPT-5.5 | 78.9 | #50 |
| 8 | GPT-5.6 Sol | 77.5 | #33 |
| 9 | Qwen 3.8 27B | 77.4 | #216 |
| 10 | Gemini 3.5 Flash | 77.4 | #106 |
| 11 | Seed 2.1 Pro | 76.7 | #142 |
| 12 | Qwen 3.7 Plus | 75.9 | #215 |
| 13 | Gemini 3.1 Pro (Preview) | 75.8 | #99 |
| 14 | GPT-5.6 Terra | 75.6 | #69 |
| 15 | Seed 2.0 Pro | 74.9 | #218 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=gamehorizon-offline-t2-multi-horizon-decomposition · How It Works · Data refreshed daily, snapshot 2026-11-18.