ModelBeat
Models/Claude Opus 4.5
All models
A

Anthropic: Claude Opus 4.5

anthropic/claude-opus-4.5

Claude Opus 4.5 is Anthropic’s proprietary AI model, released Nov 24, 2025. It costs $5/M input and $25/M output tokens, and handles a 200K-token context window. Its strongest use case is coding, where it ranks in the 72nd percentile of the models Model Beat tracks.

Claude Opus 4.5 is Anthropic’s frontier reasoning model optimized for complex software engineering, agentic workflows, and long-horizon computer use.

Context
200K
Input
$5 / 1M
Output
$25 / 1M
Released
Nov 24, 2025
Modalities
fileimagetexttext
Strongest atCoding top 28%Reasoning & Knowledge top 37%Agentic & Tools top 37%

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Claude Opus 4.5 lands among all models tracked on Model Beat.

57.0
Intelligence Index
Epoch AI
57th percentile of tracked models
72.0
Coding Index
Epoch AI
72nd percentile of tracked models
63.0
Agentic Index
Epoch AI
63rd percentile of tracked models

Reasoning

8 evals
ARC-AGI80.0%

Abstract visual reasoning

ARC-AGI-237.6%

Harder abstract reasoning

GPQA Diamond86.0%

Graduate-level scientific reasoning

Humanity's Last Exam25.2%

Frontier of human expert knowledge

MMLU-Pro89.5%

Broad expert knowledge

SimpleBench62.0%

Common-sense trick questions

SimpleQA Verified41.8%

Factual accuracy & hallucination

WeirdML63.7%

Novel ML problem-solving

Coding

5 evals
GSO (code optimization)26.5%

Code performance optimization

LiveCodeBench87.1%

Contamination-free coding

SciCode49.5%

Scientific research coding

SWE-bench Verified76.7%

Real GitHub issue resolution

WebDev Arena1512

Human-rated web development

Math

3 evals
AIME 2024/202586.1%

Olympiad-qualifier math

FrontierMath20.7%

Research-level math problems

FrontierMath Tier 44.2%

Hardest research math

Agentic & Tools

5 evals
APEX20.7%

Multi-step agentic tasks

GDPval (win/tie rate)59.6%

Economically valuable work

METR task horizon4.9 h

Autonomous task length

Terminal-Bench63.1%

Command-line agentic tasks

τ²-bench89.5%

Tool-agent-user reliability

Changelog

1
  • Claude Opus 4.5: APEX improved from 18.4% to 20.7% (+12%)benchmark · Jul 30, 2026

Compare Claude Opus 4.5 with

Frequently asked questions

What is Claude Opus 4.5?

Claude Opus 4.5 is Anthropic’s frontier reasoning model optimized for complex software engineering, agentic workflows, and long-horizon computer use.

Who created Claude Opus 4.5?

Claude Opus 4.5 is an AI model developed by Anthropic, part of the Claude family, released Nov 24, 2025.

How much does Claude Opus 4.5 cost?

Claude Opus 4.5's list price from Anthropic is $5 per million input tokens and $25 per million output tokens.

What is Claude Opus 4.5's context window?

Claude Opus 4.5 supports a context window of up to 200K tokens.

How does Claude Opus 4.5 perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Claude Opus 4.5 scores 89.5% on MMLU-Pro, 87.1% on LiveCodeBench, 59.6% on GDPval (win/tie rate).

Is Claude Opus 4.5 open source?

Claude Opus 4.5 is a proprietary model (API access), available through its provider's API rather than as a downloadable open-weight model.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Jul 30, 2026.