Models / Claude Sonnet 4.5

Anthropic

Claude Sonnet 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)

1,449

Overall · rank 20 of 44

Unit: Arena rating, higher is better
Range: 1,443 to 1,455
Sample: 31455 votes
Configuration: Claude Sonnet 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)

1,175

Overall · rank 21 of 34

Unit: Arena rating, higher is better
Range: 1,170 to 1,180
Sample: 127378 votes
Configuration: Claude Sonnet 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)

1,449

Overall · rank 54 of 177

Unit: Arena rating, higher is better
Range: 1,446 to 1,451
Sample: 77332 votes
Configuration: Claude Sonnet 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)

1,447

Overall · rank 54 of 177

Unit: Arena rating, higher is better
Range: 1,444 to 1,449
Sample: 74579 votes
Configuration: Claude Sonnet 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)

1,461

Business, management and finance · rank 33 of 402

Unit: Arena rating, higher is better
Range: 1,455 to 1,466
Sample: 15821 votes
Configuration: Claude Sonnet 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)

1,459

Business, management and finance · rank 35 of 402

Unit: Arena rating, higher is better
Range: 1,453 to 1,464
Sample: 16014 votes
Configuration: Claude Sonnet 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)

1,454

Creative writing · rank 18 of 407

Unit: Arena rating, higher is better
Range: 1,448 to 1,460
Sample: 12359 votes
Configuration: Claude Sonnet 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)

1,450

Creative writing · rank 20 of 407

Unit: Arena rating, higher is better
Range: 1,444 to 1,456
Sample: 12639 votes
Configuration: Claude Sonnet 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

Expert prompts · rank 35 of 359

Unit: Arena rating, higher is better
Range: 1,475 to 1,491
Sample: 6139 votes
Configuration: Claude Sonnet 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)

1,498

Expert prompts · rank 17 of 359

Unit: Arena rating, higher is better
Range: 1,490 to 1,506
Sample: 6061 votes
Configuration: Claude Sonnet 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)

1,462

Instruction following · rank 28 of 409

Unit: Arena rating, higher is better
Range: 1,457 to 1,466
Sample: 24284 votes
Configuration: Claude Sonnet 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)

1,464

Instruction following · rank 24 of 409

Unit: Arena rating, higher is better
Range: 1,459 to 1,468
Sample: 24600 votes
Configuration: Claude Sonnet 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)

1,455

Overall · rank 53 of 409

Unit: Arena rating, higher is better
Range: 1,452 to 1,458
Sample: 82451 votes
Configuration: Claude Sonnet 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)

1,457

Overall · rank 53 of 409

Unit: Arena rating, higher is better
Range: 1,454 to 1,459
Sample: 83689 votes
Configuration: Claude Sonnet 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)

1,454

Writing, literature and language · rank 27 of 408

Unit: Arena rating, higher is better
Range: 1,449 to 1,459
Sample: 18925 votes
Configuration: Claude Sonnet 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)

1,452

Writing, literature and language · rank 28 of 408

Unit: Arena rating, higher is better
Range: 1,447 to 1,457
Sample: 19307 votes
Configuration: Claude Sonnet 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

86.6%

Irrelevance detection · rank 26 of 109

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

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

95.0%

Irrelevance detection · rank 5 of 109

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

Reported by Berkeley Function Calling Leaderboard (BFCL) V4

65.0%

Memory · rank 2 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

5.4%

Memory · rank 89 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

61.4%

Multi-turn tasks · rank 9 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

1.6%

Multi-turn tasks · rank 97 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

73.2%

Overall accuracy · rank 2 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

24.9%

Overall accuracy · rank 89 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

68.8%

Relevance detection · rank 80 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

37.5%

Relevance detection · rank 103 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

88.7%

Single-turn calls (curated) · rank 13 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

59.8%

Single-turn calls (curated) · rank 97 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

81.1%

Single-turn calls (user-contributed) · rank 7 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

46.6%

Single-turn calls (user-contributed) · rank 102 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

81.0%

Web search · rank 6 of 109

Unit: % correct, higher is better
Configuration: claude-sonnet-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

16.0%

Web search · rank 47 of 109

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

Reported by OpenHands Index

53.0%

Average · rank 15 of 34

Unit: % resolved, higher is better
Configuration: claude-sonnet-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

36.8%

Front end (SWE-Bench Multimodal) · rank 11 of 34

Unit: % resolved, higher is better
Configuration: claude-sonnet-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

12.5%

Greenfield (Commit0) · rank 25 of 34

Unit: % resolved, higher is better
Configuration: claude-sonnet-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

72.7%

Information gathering (GAIA) · rank 10 of 34

Unit: % resolved, higher is better
Configuration: claude-sonnet-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

74.2%

Issue resolution (SWE-Bench) · rank 17 of 34

Unit: % resolved, higher is better
Configuration: claude-sonnet-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

68.8%

Testing (SWT-Bench) · rank 16 of 34

Unit: % resolved, higher is better
Configuration: claude-sonnet-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

48.0%

Consistency (pass^4) · rank 7 of 8

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

72.0%

Task success (pass^1) · rank 7 of 8

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

10.3%

Consistency (pass^4) · rank 1 of 3

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

25.3%

Task success (pass^1) · rank 2 of 3

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

39.5%

Consistency (pass^4) · rank 8 of 8

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

72.4%

Task success (pass^1) · rank 8 of 8

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

64.0%

Consistency (pass^4) · rank 6 of 8

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

84.9%

Task success (pass^1) · rank 7 of 8

Unit: % of tasks, higher is better
Configuration: claude-sonnet-4.5
Measured: 24 Feb 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

5.0%

Requested-length error · rank 34 of 370

Unit: % off the requested word count, lower is better
Configuration: claude-sonnet-4.5
Measured: 30 Sep 2025
Not shown: Not other format limits such as character counts or bullet counts.

Reported by UGI Leaderboard

8.0%

Requested-length error · rank 55 of 370

Unit: % off the requested word count, lower is better
Configuration: claude-sonnet-4.5
Measured: 30 Sep 2025
Not shown: Not other format limits such as character counts or bullet counts.

Reported by UGI Leaderboard

0.38

Style adherence · rank 71 of 370

Unit: score from 0 to 1, higher is better
Configuration: claude-sonnet-4.5
Measured: 30 Sep 2025
Not shown: Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.

Reported by UGI Leaderboard

0.41

Style adherence · rank 16 of 370

Unit: score from 0 to 1, higher is better
Configuration: claude-sonnet-4.5
Measured: 30 Sep 2025
Not shown: Not brand voice on your own examples: UGI's prompts are private and lean towards creative writing.

Reported by UGI Leaderboard

66.9

Writing score · rank 49 of 370

Unit: score out of 100, higher is better
Configuration: claude-sonnet-4.5
Measured: 30 Sep 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

64.2

Writing score · rank 73 of 370

Unit: score out of 100, higher is better
Configuration: claude-sonnet-4.5
Measured: 30 Sep 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

95.6%

Answer rate · rank 90 of 108

Unit: % of documents, higher is better
Configuration: Claude Sonnet 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

12.0%

Hallucination rate · rank 78 of 108

Unit: % of summaries, lower is better
Configuration: Claude Sonnet 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 Sonnet 4.5

Reported by others

Compare Claude Sonnet 4.5 with