Claude Sonnet 5

AnthropicReleased 30 Jun 2026ClosedFlagshipDate verified

Compare Announcement

169158–180Index score#22 of 229 models

3/ 8Core tests taken3 of 6 areas · plus 14 anchor tests

−16Versus the frontier at releasebest then: Claude Fable 5Replaced by Claude Sonnet 5.5

$2.0/ $10Price per million tokensinput / output, list price

1MContext windowtokens it reads at once

Where it sitson the frontier

GPT-5.3-Codex reached this score 5 months earlier.

Test resultsby area · tick = best by any model

Reasoning & knowledgeevidence

GPQA Diamond

91%

Humanity's Last Exam

no result

Mathsevidence

FrontierMath T1-3

66%

Codingevidence

SWE-bench Verified

no result

Terminal-Bench

no result

Agentsevidence

APEX-Agents

55%

Novel problemsno evidence yet

ARC-AGI-2

no result

Long tasksno evidence yet

METR time horizon

no result

Independent results from the Epoch AI benchmarking hub, best reasoning setting. The core tests are shown here; anchor tests also inform the score and can be shown above.

What others sayrank on each leaderboard

The Model Indexindependent tests, our method#22of 219 · 169

LMArena · Codingcoding questions in chat#35of 383 · 1519 Epoch Capabilities Indexoverall capability from many benchmarks#23of 239 · 156.3 LMArena · Mathsmaths questions in chat#42of 373 · 1474 LMArena · Textoverall chat quality, judged by people#51of 388 · 1462 LMArena · Creative writingcreative writing#56of 386 · 1436 LMArena · Visionunderstanding images#34of 149 · 1265 LMArena · WebDevbuilding web apps and front ends#37of 123 · 1540 LMArena · Documentsworking with long documents#16of 41 · 1466 LiveBench · Overalloverall, with fresh questions to limit contamination#28of 58 · 76

Further left is better. Leaderboards measure different things: LMArena is people voting blind between two answers, the others are test scores. Click any row for the source.