Claude Opus 5.5

Maker

Anthropic

Released

22 Sep 2026

Open weights

No, you use it through the maker's app or paid API

License

Owned by the maker, not open

Price

$4 for every million tokens you send, $20 for every million tokens it writesA token is a small piece of a word · checked 17 Oct 2026 · price source

Official page

Official page for Claude Opus 5.5 (anthropic.com)

• All-roundArtificial Analysis Intelligence Index

  58 index points #1 of 17 models scored independent

  One combined score from many independent tests of how smart a model is.

  setting: max with fallback · as of 17 Oct 2026 · source

• Chat & writingLMArena Text

  1508 Arena score #2 of 15 models scored independent

  When real people compare two AI answers side by side without knowing which AI wrote them, which one do they pick more often?

  setting: High · as of 15 Oct 2026 · source

• Build websites & appsLMArena WebDev

  1807 Arena score #1 of 18 models scored independent

  When people ask two AIs to build the same website or app and compare the results, which one do they pick?

  setting: Max · as of 17 Oct 2026 · source

• CodingLiveBench 2.0 Coding

  96.3 score out of 100 #2 of 15 models scored independent

  Fresh coding tasks with tests that check the answer, changed often so models cannot memorize them.

  setting: highest supported effort · as of 14 Oct 2026 · source

• ReasoningLiveBench 2.0 Reasoning

  92.2 score out of 100 #3 of 15 models scored independent

  Fresh logic and puzzle questions with one right answer, changed often so models cannot memorize them.

  setting: highest supported effort · as of 14 Oct 2026 · source

• Science questionsGPQA Diamond

  91% #2 of 6 models scored maker's own number

  Very hard science questions in biology, chemistry and physics, written by PhD experts.

  as of 17 Oct 2026 · source

• Hardest examHumanity's Last Exam (no tools)

  64.4% #1 of 6 models scored maker's own number

  2,500 expert questions written to be too hard for AI, answered without web search or code.

  as of 22 Sep 2026 · source

• New puzzlesARC-AGI-2

  93.3% #2 of 11 models scored independent

  New picture puzzles a model has never seen, so it has to reason instead of remember. Harder for AI than for people.

  setting: High, $0.408 per task · as of 18 Oct 2026 · source

• Agent tasksTerminal-Bench 4.0

  64.8 % of tasks done #1 of 6 models scored independent

  Can an AI agent finish real jobs in a computer's command line, like setting up software or fixing a broken project?

  setting: max, Claude Code · as of 18 Oct 2026 · source

• Most used (this week)OpenRouter usage (this week)

  5.95 trillion tokens this week #3 of 4 models scored independent

  How much developers actually use each model through OpenRouter, a popular service that connects apps to many AI models.

  setting: week to 16 Oct · as of 16 Oct 2026 · source

• SpeedOutput speed

  97 tokens per second #8 of 16 models scored independent

  How many tokens (pieces of words) per second the model writes once it starts answering. Higher is faster.

  setting: max with fallback · as of 17 Oct 2026 · source

• AnthropicClaude Fable 5.153 index pointsTied #2 of 17 models scored
• GoogleGemini 4 Argon53 index pointsTied #2 of 17 models scored
• OpenAIGPT-6 Astra53 index pointsTied #2 of 17 models scored

No stories about Claude Opus 5.5 on ITG Now yet.

More from Anthropic

• Anthropic · 18 OctAnthropic puts 5% of its shares in a public fund
• Anthropic · 17 OctAnthropic signs first Canada data center deal
• Anthropic · 17 OctAnthropic quietly pitches its own AI safety bill
• Anthropic · 16 OctClaude solves a 25-year-old encryption mystery