Claude Opus 4.6 is Anthropic's most intelligent model released in February 2026, built for agents that operate across entire workflows rather than single prompts. It features a 1M-token context window, 128K max output tokens, and the ability to spawn and coordinate multiple sub-agents working in parallel — a capability called Agent Teams. With adaptive thinking that dynamically adjusts reasoning depth, the model excels at large codebases, complex refactors, sustained knowledge work, and end-to-end project execution, producing near-production-ready documents and analyses in a single pass.
Anthropic ProAnthropic Max (5x)Anthropic Max (20x)API|VisionReasoningWeb Search|Proprietary Model
Knowledge Cutoff
2025-09-01
Input → Output Format
Context Memory
1MIN128KOUT
AI Performance Evaluation
Arena Overall Score
1502
±5As of 2026-05-01
Overall Rank
🥈 No.2
22,385 Votes
Arena by Ability
Hard Prompts
1536±6🥇 No.1
Expert Knowledge
1544±15🥈 No.2
Instruction Following
1518±8🥇 No.1
Conversation Memory
1515±10🥈 No.2
Creative
1493±11🥈 No.2
Coding
1554±9🥈 No.2
Math
1513±16🥈 No.2
Arena by Occupation
Creative Writing
1496±9🥈 No.2
Social Sciences
1517±10🥈 No.2
Media
1487±10🥇 No.1
Business
1502±10🥉 No.3
Healthcare
1514±15No.5
Legal
1510±15🥉 No.3
Software
1542±7🥈 No.2
Mathematics
1519±18🥈 No.2
Source:Arena Intelligence
Overall
AA Intelligence Index
53%↑14%
LiveBench
77%↑16%
ForecastBench
60%↑0%
Reasoning & Math
GPQA Diamond
90%↑7%
HLE
37%↑19%
LB Reasoning
89%↑20%
LB Math
89%↑15%
LB Data
70%↑17%
Coding
AA Coding Index
48%↑12%
LB Coding
78%↑5%
LB Agentic
62%↑17%
TAU2
92%↑12%
TerminalBench
46%↑12%
SciCode
52%↑10%
Language & Instructions
IFBench
53%↓10%
AA-LCR
71%↑9%
Hallucination (HHEM)
12%↑2%
Factual (HHEM)
88%↓2%
LB Language
83%↑11%
LB IF
63%↑12%
Output Speed
Standard Mode
45tok/s↓32
First Output 1.75s
Reasoning Mode
50tok/s↓37
First Output 11.83s
