Claude Sonnet 4
ModelActiveby Anthropic · family “claude-sonnet”
Claude Sonnet 4 significantly enhances the capabilities of its predecessor, Sonnet 3.7, excelling in both coding and reasoning tasks with improved precision and controllability. Achieving state-of-the-art performance on SWE-bench (72.7%),... (description from the OpenRouter listing)
in: imagein: textin: fileout: textReasoningTool use
Epoch Capabilities Index
141.7 #102 of 268
90% CI 138.7 – 143.0
Released
22 May 2025
Context window
200K tokens
Max output
64K tokens
Input $ / 1M tokens
$3.00
Output $ / 1M tokens
$15.0
Cached input $ / 1M
$0.30
Each benchmark shown separately with its own source
Benchmark results (20)
| Benchmark | Domain | Score | vs best recorded | Setting | Run | Source |
|---|---|---|---|---|---|---|
| FrontierMath-Tier-4-2025-07-01-Privatesuperseded | math | 0.0% | 0% | — | 1 Jul 2025 | Epoch ↗ |
| DeepResearch Bench | agents | 46.6% | 84% | 2K | — | External ↗ |
| GSO-Bench | coding | 4.9% | 10% | — | — | External ↗ |
| ARC-AGI-2 | reasoning | 5.9% | 6% | 16K | — | External ↗ |
| METR Time Horizons | agents | 62.0% | 73% | 16K | — | External ↗ |
| GeoBench | multimodal | 37.0% | 42% | — | — | External ↗ |
| FrontierMath-2025-02-28-Privatesuperseded | math | 4.1% ±1.2 | 8% | — | 4 Jul 2025 | Epoch ↗ |
| Fiction.LiveBench | long-context | 46.9% | 48% | — | — | External ↗ |
| Lech Mazur Writing | other | 81.4% | 95% | 16K | — | External ↗ |
| VPCT | multimodal | 34.0% | 37% | 32K | — | External ↗ |
| WeirdML | coding | 46.1% ±0.0 | 49% | 16K | — | External ↗ |
| Aider polyglot | coding | 61.3% | 70% | 32K | — | External ↗ |
| OTIS Mock AIME 2024-2025 | math | 71.1% ±6.8 | 71% | 32K | 22 May 2025 | Eval log ↗ |
| The Agent Company | agents | 33.1% | 77% | — | — | External ↗ |
| SimpleBench | reasoning | 45.5% | 56% | 12K | — | External ↗ |
| Cybench | coding | 35.0% | 38% | — | — | External ↗ |
| OSWorld | agents | 43.9% | 61% | agent: claude-4-sonnet-20250514 (50 steps) | — | External ↗ |
| GPQA diamond | science | 79.2% ±2.7 | 83% | 59K | 26 May 2025 | Eval log ↗ |
| MATH level 5 | math | 84.4% ±1.0 | 86% | — | 22 May 2025 | Epoch ↗ |
| ARC-AGI | reasoning | 40.0% | 41% | 16K | — | External ↗ |
Source: Epoch AI Benchmarking Hub (CC BY 4.0). “External” rows are leaderboard results Epoch collects from third parties. Best reported setting per benchmark is shown.
API list price over time
Price history
$ per 1M tokens
13 observations (OpenRouter listing + Internet Archive snapshots). History accumulates with every ingest run; a single point means no change has been observed yet.
Events
No events recorded yet.