Anthropic model product
Claude Sonnet 4 is an Anthropic large language model in the Claude 4 family with coding, reasoning, instruction-following, steerability, and extended thinking with tool use.
Updated Sep 8, 2026. Default version: Claude Sonnet 4
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Benchmark scores for Claude Sonnet 4.
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkTAU-bench Airline | Score60.00% | Rank05 | Participants23 | Percentile81.82% | EvidenceC | Evaluated |
| BenchmarkTAU-bench Retail | Score80.50% | Rank05 | Participants25 | Percentile83.33% | EvidenceC | Evaluated |
| BenchmarkTerminal-Bench | Score35.50% | Rank13 | Participants25 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkMMMLU | Score86.50% | Rank23 | Participants49 | Percentile54.17% | EvidenceC | Evaluated |
| BenchmarkMMMU | Score74.40% | Rank25 | Participants64 | Percentile61.90% | EvidenceC | Evaluated |
| BenchmarkSWE-Bench Verified | Score72.70% | Rank52 | Participants113 | Percentile54.46% | EvidenceC | Evaluated |
| BenchmarkLM Arena Vision Style Control | Score1,206.61 rating | Rank60 | Participants107 | Percentile44.34% | EvidenceA | Evaluated |
| BenchmarkLM Arena Vision | Score1,191.29 rating | Rank68 | Participants107 | Percentile36.79% | EvidenceA | Evaluated |
| BenchmarkAIME 2025 | Score70.50% | Rank92 | Participants119 | Percentile22.88% | EvidenceC | Evaluated |
| BenchmarkLM Arena Text Style Control | Score1,400.78 rating | Rank108 | Participants210 | Percentile48.80% | EvidenceA | Evaluated |
| BenchmarkGPQA | Score75.40% | Rank113 | Participants247 | Percentile54.47% | EvidenceC | Evaluated |
| BenchmarkLM Arena Text | Score1,349.97 rating | Rank137 | Participants210 | Percentile34.93% | EvidenceA | Evaluated |
Preference and agent-evaluation results for the default version.
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenavision style control | Categorycreative writing | Rank16 | Rating / score1,194.57 | Votes156 | ObservationsN/A | Result date |
| Arenavision | Categorycreative writing | Rank18 | Rating / score1,191.33 | Votes156 | ObservationsN/A | Result date |
| Arenavision style control | Categorydiagram | Rank60 | Rating / score1,227.62 | Votes227 | ObservationsN/A | Result date |
| Arenavision style control | Categoryoverall | Rank60 | Rating / score1,206.61 | Votes1,147 | ObservationsN/A | Result date |
| Arenavision style control | Categoryocr | Rank61 | Rating / score1,215.30 | Votes727 | ObservationsN/A | Result date |
| Arenavision | Categoryhomework | Rank62 | Rating / score1,212.27 | Votes232 | ObservationsN/A | Result date |
| Arenavision style control | Categoryhomework | Rank62 | Rating / score1,221.79 | Votes232 | ObservationsN/A | Result date |
| Arenavision | Categoryocr | Rank64 | Rating / score1,196.78 | Votes727 | ObservationsN/A | Result date |
| Arenavision | Categorydiagram | Rank65 | Rating / score1,199.30 | Votes227 | ObservationsN/A | Result date |
| Arenavision style control | Categoryenglish | Rank65 | Rating / score1,196.89 | Votes528 | ObservationsN/A | Result date |
| Arenavision | Categoryoverall | Rank68 | Rating / score1,191.29 | Votes1,147 | ObservationsN/A | Result date |
| Arenavision | Categoryenglish | Rank71 | Rating / score1,185.78 | Votes528 | ObservationsN/A | Result date |
| Arenatext style control | Categorylonger query | Rank73 | Rating / score1,442.04 | Votes6,697 | ObservationsN/A | Result date |
| Arenatext style control | Categorycreative writing | Rank80 | Rating / score1,395.55 | Votes4,230 | ObservationsN/A | Result date |
| Arenatext style control | Categorycoding | Rank81 | Rating / score1,473.45 | Votes6,300 | ObservationsN/A | Result date |
| Arenatext style control | Categoryinstruction following | Rank84 | Rating / score1,415.06 | Votes8,460 | ObservationsN/A | Result date |
| Arenatext style control | Categorykorean | Rank84 | Rating / score1,357.81 | Votes920 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry writing and literature and language | Rank86 | Rating / score1,397.62 | Votes7,326 | ObservationsN/A | Result date |
| Arenatext style control | Categoryjapanese | Rank87 | Rating / score1,355.11 | Votes951 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry entertainment and sports and media | Rank89 | Rating / score1,388.24 | Votes6,010 | ObservationsN/A | Result date |
| Arenatext style control | Categorymulti turn | Rank90 | Rating / score1,420.83 | Votes5,642 | ObservationsN/A | Result date |
| Arenatext style control | Categoryhard prompts english | Rank91 | Rating / score1,444.09 | Votes6,930 | ObservationsN/A | Result date |
| Arenatext style control | Categorygerman | Rank93 | Rating / score1,391.32 | Votes1,076 | ObservationsN/A | Result date |
| Arenatext style control | Categorypolish | Rank93 | Rating / score1,391.42 | Votes3,268 | ObservationsN/A | Result date |
| Arenatext style control | Categoryspanish | Rank94 | Rating / score1,405.14 | Votes831 | ObservationsN/A | Result date |
| Arenatext style control | Categoryfrench | Rank95 | Rating / score1,427.45 | Votes564 | ObservationsN/A | Result date |
| Arenatext style control | Categoryexpert | Rank96 | Rating / score1,434.88 | Votes1,662 | ObservationsN/A | Result date |
| Arenatext style control | Categoryhard prompts | Rank97 | Rating / score1,431.50 | Votes15,346 | ObservationsN/A | Result date |
| Arenatext style control | Categoryrussian | Rank99 | Rating / score1,401.68 | Votes2,415 | ObservationsN/A | Result date |
| Arenatext style control | Categoryindustry software and it services | Rank100 | Rating / score1,444.98 | Votes11,209 | ObservationsN/A | Result date |
Official vendor API pricing appears first, followed by individual provider offers.
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderPoe | Provider model IDanthropic/claude-sonnet-4 | Regionglobal | Input / 1M$2.6 | Output / 1M$13 | Context983K | Updated |
| ProviderJiekou.AI | Provider model IDclaude-sonnet-4-20250514 | Regionglobal | Input / 1M$2.7 | Output / 1M$13.5 | Context200K | Updated |
| ProviderCortecs | Provider model IDclaude-sonnet-4 | Regionglobal | Input / 1M$2.9 | Output / 1M$14.49 | Context200K | Updated |
| ProviderNanoGPT | Provider model IDclaude-sonnet-4-20250514 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context200K | Updated |
| ProviderVertex | Provider model IDclaude-sonnet-4@20250514 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context200K | Updated |
| ProviderOpenRouter | Provider model IDanthropic/claude-sonnet-4 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context1M | Updated |
| ProviderVertex (Anthropic) | Provider model IDclaude-sonnet-4@20250514 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context200K | Updated |
| ProviderMerge Gateway | Provider model IDanthropic/claude-sonnet-4-20250514 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context200K | Updated |
| ProviderDigitalOcean | Provider model IDanthropic-claude-sonnet-4 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context1M | Updated |
| ProviderFastRouter | Provider model IDanthropic/claude-sonnet-4 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context200K | Updated |
| ProviderZenMux | Provider model IDanthropic/claude-sonnet-4 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context1M | Updated |
| Provider302.AI | Provider model IDclaude-sonnet-4-20250514 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context200K | Updated |
| ProviderOpenCode Zen | Provider model IDclaude-sonnet-4 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context1M | Updated |
| ProviderVercel AI Gateway | Provider model IDanthropic/claude-sonnet-4 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context1M | Updated |
| ProviderAbacus | Provider model IDclaude-sonnet-4-20250514 | Regionglobal | Input / 1M$3 | Output / 1M$15 | Context200K | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Provider-specific output speed and catalog latency for Claude Sonnet 4. Runtime does not affect the capability score.
Provider | Output Speed | Catalog Latency | Max Input | Max Output | Updated |
|---|
| ProviderAnthropic | Output Speed42.00 tok/s | Catalog Latency0.40 s | Max Input200K | Max Output64K | Updated |
| ProviderGoogle | Output Speed42.00 tok/s | Catalog Latency0.40 s | Max Input200K | Max Output128K | Updated |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Technical details for the model's default version.
Available versions of this model. The score column identifies the version used in the overall ranking.
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionClaude Sonnet 4 | Released | LLMBoard38.71 | ParametersN/A | Context200K | Max output128K | Open weightsNo | LicenseProprietary |
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
Key information about Claude Sonnet 4 and its available data.
7. 7% on SWE-bench and supports extended thinking with tool use.
Data as of 2026-09-08.
Common questions about Claude Sonnet 4.
Claude Sonnet 4's default version was released on May 22, 2025.
No official standard PAYG price is currently available for Claude Sonnet 4. The lowest tracked third-party offer starts at $2.6 input and $13 output via Poe.
Claude Sonnet 4 was created by Anthropic.
The default version has a 200K token context window.
No. The default version is not marked as having publicly available weights.
15 provider offerings are linked to the default version.