Inference
EN
Anthropic

Claude 4 Sonnet

Claude 4 Sonnet is a standard model from Anthropic. 13 benchmarks count toward its index, in 7 categories.

  • Closed weights
  • Standard
  • Input: text
  • Released 2025-05
RankedEnough results for an index.
IndexAll categories, out of 100.39.5±5.0Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.40/sMedian 67/s
First tokenSeconds to the first output token.1.3 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$3.00Median $0.60
Output / 1MUS dollars per 1M output tokens.$15.00Median $2.55
ContextMaximum tokens in one request.200KMedian 262K
EloLMArena rating and rank.—Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.3 counted43.9166 / 218
CodingCode writing and repair.2 counted52.1126 / 209
ReasoningLogic problems and puzzles.2 counted38.2173 / 227
MultimodalTasks with images and text.1 counted36.4—
KnowledgeFacts and expert knowledge.3 counted36.0218 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted31.5—
MathMath problems.1 counted17.0—

Results

13 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
SWE-bench Lite · SWE-bench/experimentsCoding57.5%1 / 15
SWE-bench Lite · swebench.comCoding56.7%1 / 15
SWE-bench Verified · swebench.comCoding68.5%4 / 46
SWE-bench Verified · SWE-bench/experimentsCoding70.8%13 / 65
Software Engineering Benchmark VerifiedCoding72.7%57 / 114
Critical Physics TasksReasoning1.1%118 / 214
Artificial Analysis IFBenchInstruction45.4%86 / 146
τ²-Bench Tool-Agent-User EvaluationAgentic52.3%94 / 149
Artificial Analysis GPQA DiamondKnowledge68.3%135 / 199
Artificial Analysis Omniscience AccuracyKnowledge22.7%118 / 164
Gert Labs Composite Game BenchmarkAgentic39.7%38 / 52
Artificial Analysis Long Context ReasoningReasoning44.0%165 / 217
Artificial Analysis MMMU-ProMultimodal62.4%87 / 113
Artificial Analysis Humanity's Last ExamKnowledge4.3%190 / 222
JobBenchAgentic18.4%31 / 36
American Invitational Mathematics Examination 2025Math33.1%48 / 53
SWE-bench Multimodal · SWE-bench/experimentsMultimodal35.6%2 / 7

Sources

BenchLMCC BY-NC 4.0SWE-bench teamNo licence stated

History

5 releases
Index40.4 → 39.5
23 Oct 202612 Nov 2026
Output / 1M$15.00 → $15.00
23 Oct 202612 Nov 2026

Questions

How much does Claude 4 Sonnet cost?
$3.00 per 1M input tokens, $15.00 per 1M output tokens.
How good is Claude 4 Sonnet?
Its index is 39.5 out of 100. That is place 210 of 283 ranked models.
What is Claude 4 Sonnet best at?
Coding: 52.1 out of 100, place 126 of 209.
How fast is Claude 4 Sonnet?
40 output tokens per second, 1.3 s to the first token.
What is the context window of Claude 4 Sonnet?
200K tokens.
Is Claude 4 Sonnet open source?
No. Its weights are not published.
When was Claude 4 Sonnet released?
May 2025.

Badge

Claude 4 Sonnet on Inference360: place 210 of 283, index 39.5
HTML
<a href="https://inference360.ai/models/claude-4-sonnet"><img src="https://inference360.ai/badges/claude-4-sonnet.svg" alt="Claude 4 Sonnet on Inference360: place 210 of 283, index 39.5" height="20"></a>
Markdown
[![Claude 4 Sonnet on Inference360: place 210 of 283, index 39.5](https://inference360.ai/badges/claude-4-sonnet.svg)](https://inference360.ai/models/claude-4-sonnet)

Compare with

Kimi K239.6Claude 4 Opus Thinking47.0

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from Anthropic

All models
Claude 4 Opus Thinking47.0Claude 4 OpusUnrankedClaude Sonnet 443.0Claude 4.1 Opus Thinking47.0Claude 4.1 OpusUnranked