OpenAI GPT-6.1 Terra Launch - OSWorld 2.1, offline set: leaderboard

Official OpenAI October 26, 2026 GPT-6.1 Terra launch effort sweep; Score (%). Explicit version/effort/model settings are preserved in the board and row names. Shared Astra-card safety/health rows retain the unchanged GPT-6.1 Sol anchor only; changed older comparison checkpoints are not overwritten. Metric: Score (%).

Metric: Score (%). Source: openai.com. Saturation forecast: Around April 2027. 30 models tracked.

Top models

#ModelScoreOverall rank
1Claude Sonnet 5.5 (Max)83.9#13 (Claude Sonnet 5.5)
2Claude Sonnet 5.5 (xHigh)81#13 (Claude Sonnet 5.5)
3GPT-6 Astra (Max)80.9#5 (GPT-6 Astra)
4GPT-6 Astra (xHigh)79#5 (GPT-6 Astra)
5GPT-6 Astra (High)78.3#5 (GPT-6 Astra)
6GPT-6.1 Sol (Max)78#6 (GPT-6.1 Sol)
7GPT-6.1 Sol (xHigh)75.9#6 (GPT-6.1 Sol)
8GPT-6.1 Sol (High)75.5#6 (GPT-6.1 Sol)
9GPT-6 Astra (Medium)75.4#5 (GPT-6 Astra)
10GPT-6.1 Terra (Max)74.2#10 (GPT-6.1 Terra)
11GPT-6.1 Sol (Medium)73.5#6 (GPT-6.1 Sol)
12Claude Sonnet 5.5 (High)73.3#13 (Claude Sonnet 5.5)
13GPT-6.1 Terra (xHigh)71.3#10 (GPT-6.1 Terra)
14GPT-6.1 Terra (High)68.7#10 (GPT-6.1 Terra)
15GPT-6.1 Terra (Medium)66.7#10 (GPT-6.1 Terra)

No result here: #3 Claude Opus 5.5, #8 Gemini 4 Argon, #9 Claude Fable 5.1.

Interactive version: theaggregate.ai/benchmark?slug=openai-gpt-6-1-terra-launch-osworld-2-1-offline-set · How It Works · Data refreshed daily, snapshot 2026-10-27.