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%),...
Specification
anthropic listing
Context
1M
Max output
64K
Released
—
Knowledge cutoff
2025-01-31
Retires
2026-06-15
Open weights
—
Input
image, text, file
Output
text
Price
US dollars per million tokens · anthropic list price
Input
$3.00
Output
$15.00
Cache read
$0.30
Cache write
$3.75
Long-context price tiers
the rate above each context threshold
| Above | Price | $/M |
|---|---|---|
| 200K | Cache read | $0.60 |
| 200K | Cache write | $7.50 |
| 200K | Input | $6.00 |
| 200K | Output | $22.50 |
Quality
1 benchmark · 1 source
Composite
86th
percentile of 334 scored models
Rank
49 / 334
±0.10 sd
Evidence
1 × 1
too little evidence to place with confidence
Effort range
—
one configuration only
coding68th1/3 bench
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| SWE-Bench verified | 71.0 | default · SWE-agent · 8 runs | SWE-bench | — |
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 6 hosts
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| anthropicvendor | $3.00 | $15.00 | $0.30 | $3.75 | 1M | 64K | 2026-06-15 |
| replicate · replicate | $3.00 | $15.00 | — | — | — | — | — |
| snowflake | $3.00 | $15.00 | $0.30 | — | 200K | 16K | — |
| vercel_ai_gateway · vercel-ai-gateway | $3.00 | $15.00 | $0.30 | $3.75 | 200K | 64K | — |
| deepinfra | $3.30 | $16.50 | — | — | 200K | 200K | — |
| heroku · heroku | — | — | — | — | 200K | 8K | — |
More from Anthropic
most-hosted first