SO-Arena v1.0 - Overall: leaderboard

SO-Arena v1.0 overall equal-weight mean of11domain scores on the source-selected1232-task panel. Domain scores average native task credits for static decisions, tool outcomes or web progress, then scale to0–100. The full release has2919tasks; this leaderboard emphasizes the highest difficulty tiers and keeps all untiered sentiment/knowledge tasks. Metric: Equal-domain macro credit (%). Higher is better.

Metric: Equal-domain macro credit (%). Source: huggingface.co. Saturation forecast: Around February 2027. 20 models tracked. Updated 14 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-6 SolMedium82.32#26
2GPT-6 LunaMedium79.68#99
3GLiDE v1.270.21
4KnowLine-31B-Gen170.19
5Sieve-31B-v367.65
6PPLX-Decider-v1.266.7
7Kev 27B61.39
8Jev 1.14.060.43
9StartLux-v1.1-4B-Instant58.96
10StartLux-v1.1-4B-AutoThink58.63
11JevAny-27B58.37
12Clef 27B56.87
13H2O-Lightning-4B54.86
14PPLX-Decider-Flash-v1.252
15Clef flash 9.5B51.61

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=so-arena-v1-0-overall · How It Works · Data refreshed daily, snapshot 2026-11-16.