Anthropic: Claude Sonnet 5
Claude Sonnet 5 is Anthropic's model accepting text, image, and file inputs, with a 1,000,000-token context window and up to 128,000 output tokens per response. It supports tool use and reasoning, which makes it suited to multi-step workflows over large documents or codebases. The provided specifications do not confirm structured output support or open weights, so teams that depend on either should verify before committing. At $2 per million input tokens and $10 per million output tokens, it is a paid model whose output costs will dominate in generation-heavy workloads. Its blended benchmark score of 91.4 rests on only three benchmarks, so treat the ranking as indicative rather than settled. Within that set, it scores 100 on the coding index, 88 on the general index, and a weaker 77 on the agentic index. Developers prioritizing coding should shortlist it. Those building autonomous agents should test it against alternatives first.
Benchmark results
Independent, published benchmarks. Blended score 91.4 across 3 benchmarks, last refreshed 2026-11-19. How scoring works →
| Benchmark | Measures | Score |
|---|---|---|
| AI Index | broad capability composite | 88.0 |
| AI Index Coding | software engineering tasks | 100.0 |
| AI Index Agentic | multi-step tool-using tasks | 77.0 |
- Model ID
- anthropic/claude-sonnet-5
- Vendor
- anthropic
- Released
- June 2026
- Tokenizer
- Claude
- Input Modalities
- text, image, file
- Output Modalities
- text
- Max Output
- 128,000 tokens
- Tool Calling
- ✓ supported
- Structured Output
- ✓ supported
- Reasoning Mode
- ✓ supported
- Vision
- ✓ accepts images
- Audio
- no
- Moderated
- yes
What it costs in practice
Computed from the current $2.00/M input and $10.00/M output rates. Run your own numbers →
| Job | Tokens | Cost |
|---|---|---|
| Summarize a 50-page report | 30k in / 1.5k out | $0.07 |
| Classify 1,000 customer emails | 500k in / 50k out | $1.50 |
| A month of a busy support chatbot | 5M in / 2M out | $30.00 |
Category rankings
Where Anthropic: Claude Sonnet 5 places across the 56 categories it ranks in. How we rank →
| # | Category | Score |
|---|---|---|
| #10 | Function / Tool CallingAgents · of 25 ranked | 154 |
| #13 | Screenshot DebuggingVision · of 25 ranked | 143 |
| #14 | Browser AutomationAgents · of 25 ranked | 172 |
| #15 | Diagram ExtractionVision · of 25 ranked | 149 |
| #15 | Chart & Graph ReadingVision · of 25 ranked | 152 |
| #16 | Frontend Component DesignCode · of 25 ranked | 146 |
| #16 | Data AnalysisData · of 25 ranked | 171 |
| #16 | Financial AnalysisProfessional · of 25 ranked | 159 |
| #16 | Fitness CoachingPersonal · of 25 ranked | 135 |
| #17 | OCR / Document ParsingData · of 25 ranked | 146 |
| #17 | Table Extraction from PDFsData · of 25 ranked | 146 |
| #18 | SOP / Process DocsBusiness · of 25 ranked | 137 |
| #18 | Experiment DesignResearch · of 25 ranked | 157 |
| #19 | Bug FixingCode · of 25 ranked | 180 |
| #19 | Math TutoringEducation · of 25 ranked | 147 |
| #19 | Resume WritingBusiness · of 25 ranked | 117 |
| #19 | Agent WorkflowsAgents · of 25 ranked | 192 |
| #19 | Coding AgentsAgents · of 25 ranked | 192 |
| #19 | Scientific CodingResearch · of 25 ranked | 180 |
| #20 | Regex WritingCode · of 25 ranked | 135 |
| #20 | Language TranslationWriting · of 25 ranked | 135 |
| #20 | Tutoring for KidsPersonal · of 25 ranked | 135 |
| #21 | Unit Test GenerationCode · of 25 ranked | 154 |
| #21 | Marketing CopyWriting · of 25 ranked | 130 |
| #21 | Standardized Test PrepEducation · of 25 ranked | 130 |
| #22 | SQL GenerationCode · of 25 ranked | 171 |
| #22 | CI/CD PipelinesCode · of 25 ranked | 144 |
| #22 | Sales / Cold EmailBusiness · of 25 ranked | 122 |
| #22 | Job Application DraftingBusiness · of 25 ranked | 122 |
| #22 | Journaling HelperPersonal · of 25 ranked | 122 |
| #22 | Recipe GenerationPersonal · of 25 ranked | 122 |
| #23 | CSV / Spreadsheet CleanupData · of 25 ranked | 153 |
| #23 | Physics TutoringEducation · of 25 ranked | 147 |
| #24 | Code ReviewCode · of 25 ranked | 167 |
| #24 | Code RefactoringCode · of 25 ranked | 164 |
| #24 | ETL ScriptingData · of 25 ranked | 154 |
| #24 | Blog Post WritingWriting · of 25 ranked | 140 |
| #24 | Creative WritingWriting · of 25 ranked | 157 |
| #24 | Screenwriting & DialogWriting · of 25 ranked | 157 |
| #24 | Academic WritingWriting · of 25 ranked | 157 |
| #24 | Legal DraftingProfessional · of 25 ranked | 157 |
| #24 | Legal ResearchProfessional · of 25 ranked | 167 |
| #24 | Contract ReviewProfessional · of 25 ranked | 177 |
| #24 | Medical Note SummarizationProfessional · of 25 ranked | 154 |
| #24 | Scientific ResearchProfessional · of 25 ranked | 167 |
| #24 | History TutoringEducation · of 25 ranked | 144 |
| #24 | Essay GradingEducation · of 25 ranked | 167 |
| #24 | RFP ResponseBusiness · of 25 ranked | 174 |
| #24 | OKRs & Strategic PlanningBusiness · of 25 ranked | 157 |
| #24 | RAG PipelinesAgents · of 25 ranked | 164 |
| #24 | Long-Context Q&AAgents · of 25 ranked | 154 |
| #24 | Fiction CollaboratorPersonal · of 25 ranked | 167 |
| #24 | Trip PlanningPersonal · of 25 ranked | 140 |
| #24 | Math ProofsResearch · of 25 ranked | 157 |
| #24 | Literature ReviewResearch · of 25 ranked | 167 |
| #25 | Character RoleplayPersonal · of 25 ranked | 149 |
Similar models
Anthropic: Claude Haiku 5.5
Anthropic: Claude Haiku 5.5 (batch)
Anthropic: Claude Sonnet 5.5
Anthropic: Claude Sonnet 5.5 (batch)
Anthropic: Claude Opus 5.5 (batch)
Anthropic: Claude Opus 5 (batch)
Quick answers
- How much does Anthropic: Claude Sonnet 5 cost?
- $2.00 per million input tokens and $10.00 per million output tokens.
- What is Anthropic: Claude Sonnet 5's context window?
- 1,000,000 tokens, roughly 1,500 pages of text in a single request.
- Does Anthropic: Claude Sonnet 5 support tool calling?
- It supports tool calling, structured output, a reasoning mode.
- Can Anthropic: Claude Sonnet 5 process images?
- Yes, it accepts image input.