Which model leads OSWorld 2.0?

Q: Which model leads OSWorld 2.0?

Claude Opus 5.5 tops the board at 81.8

OSWorld 2.0 Leaderboard — 19 Models Ranked

Pro Mode — raw scores, one benchmark under the lensofficial card harness

The field · 18 measured

19.481.8

18 models measured, most on the official card harness. Claude Opus 5.5 tops the board at 81.8.

18 measured·1 new this week·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 18Bars show 19.4–81.8 · ◆ solid = first-party or better · outlined = arm's-length

01Claude Opus 5.5Anthropic81.8$4.0/M$20.0/M$0.147

02Claude Fable 5.1Anthropic77.9$10.0/M$50.0/M$0.385

03GPT-6 AstraOpenAI72.6$10.0/M$50.0/M$0.413

04GPT-6.1 SolOpenAI71.42$2.0/M$10.0/M$0.084

05Claude Opus 5+2 altsAnthropic70.6$5.0/M$25.0/M$0.212

06Gemini 4 ArgonNEW+1 altGoogle69.2$2.0/M$10.0/M$0.087

07Muse Spark 1.3Meta66.9$1.3/M$4.3/M$0.041

08Claude Fable 5Anthropic66.1$10.0/M$50.0/M$0.454

09GPT-6 SolOpenAI64.4$2.0/M$10.0/M$0.093

10GPT-5.6 Sol+1 altOpenAI62.6$4.0/M$20.0/M$0.192

11Gemini 3.8 FlashGoogle59$0.75/M$3.8/M$0.038

12Claude Opus 4.8Anthropic55.7$5.0/M$25.0/M$0.269

13GPT-6 LunaOpenAI52.7$0.10/M$0.50/M$0.006

14GPT-5.6 TerraOpenAI50.2$2.0/M$12.0/M$0.139

15Gemini 3.7 Flash+1 altGoogle47.9$0.75/M$3.8/M$0.047

16GPT-5.6 LunaOpenAI45.6$0.20/M$1.2/M$0.015

17Gemini 3.6 FlashGoogle33.8$0.75/M$3.8/M$0.067

18Qwen3.8 Flash NextAlibaba19.4$0.15/M$0.47/M$0.016

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.

Verification: 18 models scored · 0 independently verified · 18 vendor-reported · 0 with source disagreement. How these tiers are assigned