supermarket

Vision-language

Claude Opus 4.5

By Anthropic, in the Claude Opus family.

Released
24 Nov 2025
Status
Legacy
Context
200K
Max output
64K
Input per 1M tokens
$5.00
Output per 1M tokens
$25.00
Open weights
No
License
proprietary
Parameters
—
Knowledge cutoff
May 2025
Reasoning
Yes
Modalities
text, image → text

Category scores

Out of 100, from independent and third-party results only, by method category-2026-09-29.1.

Category scores
CategoryScoreRankCoverage
Knowledge & reasoning32.1524 of 4
Coding & software engineering26.5Not ranked1 of 5
Agentic tool use49.132 of 2
Vision & multimodal understanding28.3143 of 3
Factuality & hallucination45.7311 of 1
Human preference59.673 of 5

Benchmark results

Every result we hold for this model, with who ran it and where it was published.

Knowledge & reasoning

Knowledge & reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Chess PuzzlesPuzzles solved12.0%Independent32K8 Dec 2025Epoch AI
Chess PuzzlesPuzzles solved4.0%Independent—6 Aug 2026Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)89.9%Third partyhigh—https://conceptualreasoning.ai/dtbench via Epoch AI
Earthborne Rangers (EBR-bench)Objectives reached14.3%Independent128K25 Jun 2026Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.4446Third partyhigh—https://conceptualreasoning.ai/lmca via Epoch AI
Mystery Game PuzzlesPuzzles solved22.0%Independent48K25 Jul 2026Epoch AI
SimpleBenchScore (AVG@5)62.0%Third party——SimpleBench Leaderboard via Epoch AI
SimpleBenchScore (AVG@5)62.0%Third partyunknown—SimpleBench Leaderboard via Epoch AI

Scientific reasoning

Scientific reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GPQA DiamondAccuracy86.0%Independent32K24 Nov 2025Epoch AI
GPQA DiamondAccuracy85.5%Independent16K25 Nov 2025Epoch AI
GPQA DiamondAccuracy80.7%Independent—24 Nov 2025Epoch AI

Mathematics

Mathematics results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
FrontierMath Tier 4Accuracy4.9%Independent32K11 Jun 2026Epoch AI
FrontierMath Tier 4 (v1)Accuracy4.2%Independent32K25 Nov 2025Epoch AI
FrontierMath Tier 4 (v1)Accuracy4.2%Independent—25 Nov 2025Epoch AI
FrontierMath Tier 4 (v1)Accuracy2.1%Independent16K25 Nov 2025Epoch AI
FrontierMath Tiers 1-3Accuracy34.4%Independent32K11 Jun 2026Epoch AI
FrontierMath Tiers 1-3 (v1)Accuracy20.7%Independent32K25 Nov 2025Epoch AI
FrontierMath Tiers 1-3 (v1)Accuracy20.7%Independent—25 Nov 2025Epoch AI
FrontierMath Tiers 1-3 (v1)Accuracy20.3%Independent16K25 Nov 2025Epoch AI
OTIS Mock AIME 2024-2025Accuracy86.1%Independent32K24 Nov 2025Epoch AI
OTIS Mock AIME 2024-2025Accuracy81.7%Independent16K24 Nov 2025Epoch AI
OTIS Mock AIME 2024-2025Accuracy48.1%Independent—24 Nov 2025Epoch AI

Coding & software engineering

Coding & software engineering results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
CybenchUnguided challenges solved82.0%Self-reportedunknown—Opus 4.5 System Card via Epoch AI
GSOOPT@126.5%Third party——GSO Leaderboard via Epoch AI
GSOOPT@126.5%Third partyunknown—GSO Leaderboard via Epoch AI
METR Time HorizonsAverage task success (Epoch's column)75.0%Third party16K—METR - Measuring AI Ability to Complete Long Tasks via Epoch AI
METR Time HorizonsAverage task success (Epoch's column)73.0%Source not stated——Epoch AI via Epoch AI
SWE-bench Verified% resolved76.7%Independent—5 Feb 2026Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)63.1%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)63.1%Third party——Terminal-Bench v2 Leaderboard via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)59.1%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)59.1%Third party128K—Terminal-Bench v2 Leaderboard via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)58.4%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)57.8%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)54.3%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)52.1%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)51.9%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)51.7%Third partyunknown—https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
WeirdML (v2)Average accuracy63.7%Third party16K—https://htihle.github.io/weirdml.html via Epoch AI

Agentic tool use

Agentic tool use results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
BALROGAverage progress43.5%Third party——Balrog Leaderboard via Epoch AI
BALROGAverage progress43.0%Third partyunknown—https://balrogai.com/ via Epoch AI
BALROGAverage progress43.0%Third party64K—Balrog Leaderboard via Epoch AI
Deep Research Bench (FutureSearch)Average score54.8%Third partyhigh—https://evals.futuresearch.ai/#drb via Epoch AI
Deep Research Bench (FutureSearch)Average score53.7%Third partylow—https://evals.futuresearch.ai/#drb via Epoch AI

Computer & browser use

Computer & browser use results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
OSWorld-Verified% tasks completed66.3%Self-reported——Anthropic annoucement via Epoch AI

Professional knowledge work

Professional knowledge work results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GDPvalWins vs experts45.5%Source not stated——Epoch AI via Epoch AI

Long context

Long context results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
CL-benchTasks solved21.1%Source not statedunknown—Epoch AI via Epoch AI

Vision & multimodal understanding

Vision & multimodal understanding results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Furniture AssemblyScore28.3%Independent64K10 Sep 2026Epoch AI
GeoBenchCountry accuracy75.0%Third party——GeoBench leaderboard via Epoch AI
Visual Physics Comprehension Test (VPCT)Correct40.0%Third party32K—Tweet from VPCT creator via Epoch AI

Factuality & hallucination

Factuality & hallucination results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
SimpleQA VerifiedAccuracy (% correct)45.7%Independent32K27 Aug 2026Epoch AI

Human preference

Human preference results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Code Arena: WebDev (arena.ai)Arena score (Elo scale)1468Independent—25 Sep 2026Arena
Code Arena: WebDev (arena.ai)Votes13430Independent—25 Sep 2026Arena
Document Arena (arena.ai)Arena score (Elo scale)1465Independent—13 Sep 2026Arena
Document Arena (arena.ai)Votes7981Independent—13 Sep 2026Arena
Text Arena (arena.ai)Arena score (Elo scale)1470Independent—25 Sep 2026Arena
Text Arena (arena.ai)Votes72766Independent—25 Sep 2026Arena

Other Claude Opus versions

Sources