Compare Start here

AnthropicLanguage modelSuperseded

Claude Opus 5

The first Claude 5 Opus: a big jump in computer use and abstract reasoning.

Compare head-to-head Race it in 3D

Overall Intelligence

#8

of 56
3.5 pts behind #1

0

races led

16

weeks at #1

85.2

score

Released: Jul 24, 2026; 3 mo ago
Context: 1M; 128K max output
Price: $5 / $25; per 1M in/out · $10.00 blended
Speed: 72 tok/s; 5.5s to first token
Modalities: text · image; outputs text
Weights: Closed; API and apps

The 30-second version

The model that brought Anthropic's family onto a new generation with a standard 1M-token context, superseded two months later by Opus 5.5.

Key takeaways

1Made 1M context the default for Opus.
2Cut price versus Opus 4.x.
3Still a strong model; Opus 5.5 is better in every race.

Where it ranks

Current snapshot · illustrative scores · click any race to watch it

Opus 5Best on each axis

Overall

8Overall+3.5 pts

Coding

7Coding+4.2 pts 6Frontend & UI+19 Elo 7Backend & APIs+3.9 pts 6Distributed Systems+5.7 pts 7Full-Stack Apps+4.8 pts 6Debugging & Repair+5.1 pts 6Agentic Coding+5.4 pts 4Code Review & Security+3.3 pts 10Data Science & SQL+3.9 pts 6Mobile Apps+3.3 pts 6Terminal & DevOps+6.2 pts 13Competitive Programming+122

Hard Reasoning

9Hard Reasoning+4.7 pts 13Frontier Questions+11.1 pts 11Abstract & Novel Reasoning+7.7 pts 10Logic & Puzzles+3.2 pts 5Planning & Strategy+4.0 pts

Math

12Math+6.0 pts 13Competition Math+5.3 pts 12Research Math+8.5 pts 9Applied & Quantitative+4.1 pts

Science

7Science+4.6 pts 8Graduate Science QA+2.6 pts 7Biology & Medicine+3.8 pts 12Physics & Chemistry+4.7 pts 7Research Assistance+4.9 pts

Writing & Language

6Writing & Language+3.3 pts 4Creative Writing+18 Elo 6Technical Writing+3.2 pts 4Editing & Style+2.3 pts 8Multilingual+2.9 pts 11Instruction Following+4.8 pts 5Conversation & EQ+20 Elo

Multimodal

20Multimodal+16.3 pts 10Image Understanding+6.3 pts 12Documents & Charts+7.4 pts 20Video Understanding+18.0 pts 26Speech & Audio+26.8 pts

Tool Use & Agents

7Tool Use & Agents+4.0 pts 7Tool Use+3.6 pts 6Computer Use+6.1 pts 7Web Research+6.1 pts 6Long-Horizon+12.9 hrs

Long Context

8Long Context+5.9 pts 71M Recall+5.4 pts 9Long Reasoning+4.5 pts

Speed & Cost

38Speed & Cost+45.7 pts 36Output Speed+450 tok/s 40Latency+5.18s 34Cost-Efficiency+48.2 pts

Trust & Factuality

7Trust & Factuality+3.8 pts 6Factuality+3.1 pts 6Honesty & Calibration+3.5 pts

Overall Intelligence over time

#1 frontierOpus 5Sonnet 5.5Grok 4.7

Update history

The distilled version first; open any release for the deeper breakdown.

The Claude Opus family

Opus 4May 2025 Opus 4.1August 2025 Opus 4.5November 2025 Opus 4.6February 2026 Opus 4.8May 2026 Opus 5July 2026 Opus 5.5September 2026

60-second briefing

Opus 5, explained

AI video summary · placeholder render from the auto-written script

From Anthropic's own guide, distilled

Prompt engineering and Claude best practices

Be explicit about the outcome you want; newer Claude models follow instructions literally rather than guessing at hidden intent.
Explain why a rule exists. Context about the goal lets the model generalise the rule to cases you did not list.
Wrap documents, examples and data in XML-style tags so instructions and material never blur together.
Put long documents at the top of the prompt and your question at the end; recall and grounding both improve.
For agentic tasks, state when to act versus when to ask, and ask for a short plan before large edits.

All official guides, distilled →

About Anthropic

Safety-focused lab building the Claude family of models.

Optimising for

Reliable long-horizon agentic work (especially coding and computer use), honest and calibrated answers, and strong writing, with safety research shipped alongside capability.

San Francisco, US · founded 2021

Closest rivals

Opus 5 is 8th in overall intelligence; leader is Claude Opus 5.5.