GameHorizon - Online Short-Horizon Subtasks: leaderboard

Metric: Subtask success rate (%). Source: gamehorizon-suite.github.io. Saturation forecast: Around February 2027. 12 models tracked. Updated 18 Nov 2026.

Top models

#ModelScoreOverall rank
1GPT-6 Astra66.1#7
2Gemini 3.6 Flash56.5#77
3Kimi K346.8#60
4GPT-5.6 Terra37.1#69
5GPT-5.6 Luna33.9#129
6MiniMax-M329#277
7GLM-5V Turbo27.4#362
8Qwen 3.5 397B A17B19.4#254
9Step3 VL 10B14.5#1084
10Qwen 3.6 35B A3B11.3#480
11InternVL3.5-8B3.2#1598
12uitars-1.5-7B1.6#7710

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=gamehorizon-online-short-horizon-subtasks · How It Works · Data refreshed daily, snapshot 2026-11-18.