| Spec | Claude Opus 4 | Claude Sonnet 4 |
|---|
| Provider | Anthropic | Anthropic |
| Release Date | May 2025 | May 2025 |
| Knowledge Cutoff | 2025-01-31 | 2025-01-31 |
| Parameters | unknown | unknown |
| Context Window | 200K | 1M |
| Max Output | 32K | 64K |
| Open Source | No | No |
| License | proprietary | proprietary |
| Tokenizer | Claude | Claude |
| Modality | text+image+file->text | text+image+file->text |
| Reasoning Model | No | No |
| Moderated | No | Yes |
| Price (per 1M tokens) | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| Input | $15.00 | $3.00 |
| Output | $75.00 | $15.00 |
| Cache Read | $1.50 | $0.30 |
| Cache Write | $18.75 | $3.75 |
| Feature | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| Text Input | ||
| Image Input (Vision) | ||
| Audio Input | ||
| Video Input | ||
| File Input | ||
| Image Output | ||
| Audio Output | ||
| Tool Use | ||
| Structured Output (JSON) | ||
| Streaming | ||
| Reasoning Tokens | ||
| Web Search | ||
| Temperature Control | ||
| Top-P Sampling | ||
| Stop Sequences | ||
| Seed (Reproducibility) | ||
| Log Probabilities |
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| IFEval | 91.5 | 90.1 |
| MMLU-Pro | 83.5 | 80.2 |
| Chatbot Arena Elo | 1365 | 1310 |
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| HumanEval | 95.1 | 93.8 |
| SWE-bench Verified | 55.8 | 53.2 |
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| GPQA Diamond | 74.8 | 70.5 |
| BigBench-Hard | 93.2 | 90.8 |
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| MMMU | 74.2 | 72.5 |