Atlas

Models

← All models

Opus 4.1

Anthropic · 2025-08-05

Opus 4.1 is Anthropic's Opus-tier hybrid-reasoning model, released as an upgrade to Opus 4. It improves agentic tasks and real-world coding—including multi-file refactoring precision—as well as in-depth research and data analysis.

Benchmark scores

BenchmarkScore
A-Fantasia35.3
A-Fantasia (Valid Attempts with Retries)35.3
A-Fantasia Backwards Spelling15.0
A-Fantasia Chess19.0
A-Fantasia Chess (Valid Attempts with Retries)19.0
A-Fantasia Cube (Valid Attempts with Retries)72.0
A-Fantasia Cube Rotation72.0
A-Fantasia Spelling (Valid Attempts with Retries)15.0
AA-LCR66.3
AA-LCR v1.176.0
AIME 2024 (avg@8)83.4
AIME 2024-202578.2
AIME 2025 (Artificial Analysis avg@10)80.3
AIME 2025 (avg@8)72.9
ALE-Bench1183
AlgoTune1.3
Arena Code WebDev — Brand & Marketing1369
Arena Code WebDev — Consumer Product1378
Arena Code WebDev — Content Creation Tools1376
Arena Code WebDev — Data & Analytics1367
Arena Code WebDev — Frontend1389
Arena Code WebDev — Gaming1422
Arena Code WebDev — HTML1399
Arena Code WebDev — Reference-Based Design1414
Arena Code WebDev — Simulations1376
Arena Text — Business, Management & Financial Ops (No Style Control)1405
Arena Text — Chinese (No Style Control)1427
Arena Text — Creative Writing (No Style Control)1410
Arena Text — English (No Style Control)1426
Arena Text — Entertainment, Sports & Media (No Style Control)1398
Loading Atlas data…