Atlas

Models

← All models

Opus 4.6

Anthropic · 2026-02-05

Claude Opus 4.6 is Anthropic's February 2026 hybrid-reasoning large language model for advanced knowledge work, coding, and agentic tasks. It introduced adaptive thinking, allowing the model to decide when deeper reasoning is useful, while retaining optional manual extended thinking. It supports a 1M-token context window.

Benchmark scores

BenchmarkScore
A-Fantasia30.0
A-Fantasia Backwards Spelling17.0
A-Fantasia Chess20.0
A-Fantasia Cube Rotation53.0
AA-LCR70.7
AA-Omniscience Index13.5
Agent Red Teaming k=100 (CAIS)49.5
AIME 2024 (avg@8)97.5
AIME 2024-202595.6
AIME 202599.8
AIME 2025 (avg@8)93.8
AIME 202696.7
ALE-Bench1607
AlgoTune1.5
Antidote: Everyday Edition1039
Apex (MathArena)34.5
APEX (Mercor)14.7
Apex Shortlist86.7
APEX-Agents32.4
APEX-Agents-AA33.0
APEX-SWE40.5
APEX-v1-extended65.7
ARC-AGI-194.0
ARC-AGI-269.2
ARC-AGI-2 Semi-Private (CAISI mean)63.0
ARC-AGI-3 (Semi-Private)0.5
Arena Code WebDev — Brand & Marketing1536
Arena Code WebDev — Consumer Product1533
Arena Code WebDev — Content Creation Tools1542
Arena Code WebDev — Data & Analytics1517
Loading Atlas data…