Models / Claude Opus 4.5

Anthropic

Claude Opus 4.5

54 published results from 6 sources. Each card shows where the number comes from and what it does not measure. Results are never combined into one score.

Provider: Anthropic
Sources: 6
Our benchmarks: 0
Price: Not yet published

Reported by others

Reported by Arena (formerly LMArena)

1,465

Overall · rank 11 of 44

Unit: Arena rating, higher is better
Range: 1,454 to 1,475
Sample: 7981 votes
Configuration: Claude Opus 4.5
Measured: 13 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,187

Overall · rank 11 of 34

Unit: Arena rating, higher is better
Range: 1,181 to 1,193
Sample: 61573 votes
Configuration: Claude Opus 4.5
Measured: 24 Aug 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,459

Overall · rank 35 of 177

Unit: Arena rating, higher is better
Range: 1,456 to 1,462
Sample: 71633 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not an error rate: claims that cannot be checked on the web are skipped, and preference still carries most of the weight.

Reported by Arena (formerly LMArena)

1,459

Overall · rank 32 of 177

Unit: Arena rating, higher is better
Range: 1,456 to 1,463
Sample: 36596 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not an error rate: claims that cannot be checked on the web are skipped, and preference still carries most of the weight.

Reported by Arena (formerly LMArena)

1,472

Business, management and finance · rank 15 of 402

Unit: Arena rating, higher is better
Range: 1,466 to 1,478
Sample: 14086 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,467

Business, management and finance · rank 20 of 402

Unit: Arena rating, higher is better
Range: 1,460 to 1,475
Sample: 7034 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,461

Creative writing · rank 12 of 407

Unit: Arena rating, higher is better
Range: 1,455 to 1,468
Sample: 11375 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,469

Creative writing · rank 6 of 407

Unit: Arena rating, higher is better
Range: 1,461 to 1,477
Sample: 5680 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,503

Expert prompts · rank 13 of 359

Unit: Arena rating, higher is better
Range: 1,494 to 1,511
Sample: 5682 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,503

Expert prompts · rank 13 of 359

Unit: Arena rating, higher is better
Range: 1,491 to 1,515
Sample: 2424 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,475

Instruction following · rank 13 of 409

Unit: Arena rating, higher is better
Range: 1,470 to 1,480
Sample: 21786 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,483

Instruction following · rank 7 of 409

Unit: Arena rating, higher is better
Range: 1,477 to 1,490
Sample: 9875 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,470

Overall · rank 27 of 409

Unit: Arena rating, higher is better
Range: 1,467 to 1,473
Sample: 72766 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,474

Overall · rank 21 of 409

Unit: Arena rating, higher is better
Range: 1,470 to 1,477
Sample: 37468 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,467

Writing, literature and language · rank 15 of 408

Unit: Arena rating, higher is better
Range: 1,461 to 1,472
Sample: 16876 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Arena (formerly LMArena)

1,466

Writing, literature and language · rank 14 of 408

Unit: Arena rating, higher is better
Range: 1,459 to 1,473
Sample: 8564 votes
Configuration: Claude Opus 4.5
Measured: 25 Sep 2026
Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

84.7%

Irrelevance detection · rank 35 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not general refusal or safety behaviour.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

90.8%

Irrelevance detection · rank 14 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not general refusal or safety behaviour.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

73.8%

Memory · rank 1 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not long-term personal memory in a product: sessions are BFCL's scripted ones.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

1.9%

Memory · rank 102 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not long-term personal memory in a product: sessions are BFCL's scripted ones.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

68.4%

Multi-turn tasks · rank 4 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not open-ended agent work: the tools and tasks are BFCL's simulated APIs.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

16.1%

Multi-turn tasks · rank 56 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not open-ended agent work: the tools and tasks are BFCL's simulated APIs.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

77.5%

Overall accuracy · rank 1 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not a neutral average: the weighting is BFCL's. Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

33.5%

Overall accuracy · rank 57 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not a neutral average: the weighting is BFCL's. Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

62.5%

Relevance detection · rank 91 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not whether the call itself was right; only that one was attempted.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

68.8%

Relevance detection · rank 80 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not whether the call itself was right; only that one was attempted.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

88.6%

Single-turn calls (curated) · rank 15 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

89.7%

Single-turn calls (curated) · rank 5 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

79.8%

Single-turn calls (user-contributed) · rank 14 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

76.0%

Single-turn calls (user-contributed) · rank 35 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not reliability on your own tools: BFCL's functions and queries are a fixed test set, and a correct call is judged by its form, not by what it achieved.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

84.5%

Web search · rank 1 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not general research quality: questions have short, checkable answers.

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

13.0%

Web search · rank 52 of 109

Unit: % correct, higher is better
Configuration: claude-opus-4.5
Measured: 16 Dec 2025
Not shown: Not general research quality: questions have short, checkable answers.

Reported by OpenHands Index

60.6%

Average · rank 8 of 34

Unit: % resolved, higher is better
Configuration: claude-opus-4.5
Measured: 30 Jun 2026
Not shown: Not the model alone: the score is for the model inside the OpenHands agent, at the SDK version shown.

Reported by OpenHands Index

41.2%

Front end (SWE-Bench Multimodal) · rank 6 of 34

Unit: % resolved, higher is better
Configuration: claude-opus-4.5
Measured: 30 Jun 2026
Not shown: Not the model alone: the score is for the model inside the OpenHands agent, at the SDK version shown.

Reported by OpenHands Index

37.5%

Greenfield (Commit0) · rank 10 of 34

Unit: % resolved, higher is better
Configuration: claude-opus-4.5
Measured: 30 Jun 2026
Not shown: Not the model alone: the score is for the model inside the OpenHands agent, at the SDK version shown.

Reported by OpenHands Index

69.1%

Information gathering (GAIA) · rank 13 of 34

Unit: % resolved, higher is better
Configuration: claude-opus-4.5
Measured: 30 Jun 2026
Not shown: Not the model alone: the score is for the model inside the OpenHands agent, at the SDK version shown.

Reported by OpenHands Index

76.6%

Issue resolution (SWE-Bench) · rank 8 of 34

Unit: % resolved, higher is better
Configuration: claude-opus-4.5
Measured: 30 Jun 2026
Not shown: Not the model alone: the score is for the model inside the OpenHands agent, at the SDK version shown.

Reported by OpenHands Index

78.5%

Testing (SWT-Bench) · rank 7 of 34

Unit: % resolved, higher is better
Configuration: claude-opus-4.5
Measured: 30 Jun 2026
Not shown: Not the model alone: the score is for the model inside the OpenHands agent, at the SDK version shown.

Reported by tau2-bench

70.0%

Consistency (pass^4) · rank 2 of 8

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by tau2-bench

84.0%

Task success (pass^1) · rank 1 of 8

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by tau2-bench

11.3%

Consistency (pass^4) · rank 15 of 21

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by tau2-bench

24.7%

Task success (pass^1) · rank 17 of 21

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by tau2-bench

51.8%

Consistency (pass^4) · rank 2 of 8

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by tau2-bench

79.6%

Task success (pass^1) · rank 3 of 8

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by tau2-bench

78.1%

Consistency (pass^4) · rank 2 of 8

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by tau2-bench

92.3%

Task success (pass^1) · rank 2 of 8

Unit: % of tasks, higher is better
Configuration: claude-opus-4.5
Measured: 5 May 2026
Not shown: Not your policies or customers: the tasks and policies are tau2's own, and the customer is simulated by gpt-5.2.

Reported by UGI Leaderboard

3.0%

Requested-length error · rank 18 of 370

Unit: % off the requested word count, lower is better
Configuration: claude-opus-4.5
Measured: 24 Nov 2025
Not shown: Not other format limits such as character counts or bullet counts.

Reported by UGI Leaderboard

4.0%

Requested-length error · rank 28 of 370

Unit: % off the requested word count, lower is better
Configuration: claude-opus-4.5
Measured: 24 Nov 2025
Not shown: Not other format limits such as character counts or bullet counts.

Reported by UGI Leaderboard

0.37

Style adherence · rank 108 of 370

Unit: score from 0 to 1, higher is better
Configuration: claude-opus-4.5
Measured: 24 Nov 2025
Not shown: Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.

Reported by UGI Leaderboard

0.39

Style adherence · rank 40 of 370

Unit: score from 0 to 1, higher is better
Configuration: claude-opus-4.5
Measured: 24 Nov 2025
Not shown: Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.

Reported by UGI Leaderboard

70.4

Writing score · rank 19 of 370

Unit: score out of 100, higher is better
Configuration: claude-opus-4.5
Measured: 24 Nov 2025
Not shown: Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.

Reported by UGI Leaderboard

70.3

Writing score · rank 20 of 370

Unit: score out of 100, higher is better
Configuration: claude-opus-4.5
Measured: 24 Nov 2025
Not shown: Not business copy quality: UGI's prompts are private and lean towards creative writing, and models that often refuse get no score.

Reported by Vectara Hallucination Leaderboard

98.7%

Answer rate · rank 74 of 108

Unit: % of documents, higher is better
Configuration: Claude Opus 4.5
Measured: 22 Sep 2026
Not shown: Not a quality score: a low rate usually means content filters were triggered, and hallucination rates are measured on answered documents only.

Reported by Vectara Hallucination Leaderboard

10.9%

Hallucination rate · rank 70 of 108

Unit: % of summaries, lower is better
Configuration: Claude Opus 4.5
Measured: 22 Sep 2026
Not shown: Not errors in open questions or other tasks: only summarisation, judged by Vectara's own model (HHEM-2.3), not by people, on news-style documents rather than your data.

Claude Opus 4.5

Reported by others

Compare Claude Opus 4.5 with