Atlas

Models

← All models

Opus 4.8

Anthropic · 2026-05-28

Opus 4.8 was released as Anthropic's most capable general-access model, improving software engineering, agentic tool use, and knowledge work over Opus 4.7. It supports optional adaptive thinking and is a permitted fallback target for refused Fable 5 requests.

Benchmark scores

BenchmarkScore
A-Fantasia50.0
A-Fantasia Backwards Spelling28.0
A-Fantasia Chess31.0
A-Fantasia Cube Rotation91.0
AA-LCR67.7
AA-Omniscience Index27.4
ActiveVision2.4
ActiveVision with tools24.7
Agent Red Teaming k=100 (CAIS)47.0
Agents' Last Exam (ALE)27.0
Agents' Last Exam ALE-CLI Pass Rate25.7
Agents' Last Exam ALE-CLI Score44.6
Agents' Last Exam Full-Spectrum Pass Rate23.6
Agents' Last Exam Full-Spectrum Score43.1
Agents' Last Exam Last-Exam Pass Rate2.6
Agents' Last Exam Last-Exam Score14.4
Agents' Last Exam Near-Term Pass Rate43.3
Agents' Last Exam Near-Term Score64.0
Agents' Last Exam Overall Score45.2
Agents' Last Exam Unlicensed Full-Spectrum Pass Rate24.0
Agents' Last Exam Unlicensed Full-Spectrum Score41.9
Agents' Last Exam Unlicensed Last-Exam Pass Rate3.3
Agents' Last Exam Unlicensed Last-Exam Score16.1
Agents' Last Exam Unlicensed Near-Term Pass Rate43.3
Agents' Last Exam Unlicensed Near-Term Score64.0
Agents' Last Exam Unlicensed Overall Pass Rate29.1
Agents' Last Exam Unlicensed Overall Score46.9
AHK-Eval100.0
AHK-Eval Algorithms83.3
AHK-Eval Data Structures66.7
Loading Atlas data…