SO-Arena v1.0 - Text-based Web Interaction: leaderboard

SO-Arena v1.0 Text-based Web Interaction domain score 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: Native panel task credit mean (%). Higher is better.

Metric: Native panel task credit mean (%). Source: huggingface.co. Saturation forecast: Around May 2027. 20 models tracked. Updated 14 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-6 SolMedium53.22#26
2PPLX-Decider-v1.250.24
3GLiDE v1.248.96
4GPT-6 LunaMedium47.5#99
5Kev 27B45.43
6KnowLine-31B-Gen144.03
7Sieve-31B-v342.82
8Clef 27B42.57
9JevAny-27B41.3
10Jev 1.14.040.66
11PPLX-Decider-Flash-v1.233.65
12StartLux-v1.1-4B-Instant31.47
13StartLux-v1.1-4B-AutoThink30.61
14H2O-Lightning-4B28.25
15Clef flash 9.5B25.18

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-text-based-web-interaction · How It Works · Data refreshed daily, snapshot 2026-11-16.