Atlas

Models

← All models

Claude Opus 5.5

Anthropic · 2026-09-22

Claude Opus 5.5 is Anthropic's September 22, 2026 model for long-running coding and knowledge work. Adaptive thinking is always on, with low, medium, high, xhigh, and max effort and medium as the default. Production safeguards can route flagged requests to fallback models. The published knowledge cutoff is June 2026, without day precision.

Benchmark scores

BenchmarkScore
AA-3D v1.0 Elo1196
AA-AnalystAgent (pass@1)64.3
AA-AnalystAgent (pass@5)75.0
AA-AnalystAgent (pass^5)56.3
AA-Briefcase v1.1 Elo1800
AA-LCR v1.184.7
AA-Omniscience Index46.4
AA-Robotics v0.148.8
ActiveVision41.2
ActiveVision with tools82.4
AdvancedIF82.8
AdvancedIF (rubric mean score)93.5
Agent Red Teaming k=100 (CAIS)24.4
Agents' Last Exam (ALE)38.2
Agents' Last Exam ALE-CLI Pass Rate34.3
Agents' Last Exam ALE-CLI Score63.7
Agents' Last Exam Full-Spectrum Pass Rate32.7
Agents' Last Exam Full-Spectrum Score58.4
Agents' Last Exam Last-Exam Pass Rate15.8
Agents' Last Exam Last-Exam Score33.9
Agents' Last Exam Near-Term Pass Rate53.7
Agents' Last Exam Near-Term Score83.9
Agents' Last Exam Overall Score63.2
Agents' Last Exam Unlicensed Full-Spectrum Pass Rate34.0
Agents' Last Exam Unlicensed Full-Spectrum Score58.5
Agents' Last Exam Unlicensed Last-Exam Pass Rate20.0
Agents' Last Exam Unlicensed Last-Exam Score36.4
Agents' Last Exam Unlicensed Near-Term Pass Rate53.7
Agents' Last Exam Unlicensed Near-Term Score83.9
Agents' Last Exam Unlicensed Overall Pass Rate41.1
Loading Atlas data…