Flagship Claude model for deep reasoning, coding, and long-horizon agents
Specification
most-agreed values
Context
200K
Max output
200K
Released
2025-11-25
Knowledge cutoff
2025-11
Retires
—
Open weights
no
Input
text, image
Output
text
Price
US dollars per million tokens · most-agreed
Input
$5.00
Output
$25.00
Cache read
$0.50
Cache write
$6.25
Quality
1 benchmark · 1 source
Composite
92nd
percentile of 334 scored models
Rank
29 / 334
±0.10 sd
Evidence
1 × 1
too little evidence to place with confidence
Effort range
2.4
points between effort settings
coding87th1/3 bench
By reasoning effort
every setting placed on the same scale as the leaderboard
| Setting | Composite | Benchmarks behind it |
|---|---|---|
| high | 92nd | 1 |
| medium | 92nd | 1 |
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| SWE-Bench verified | 79.2 | default · Sonar Foundation Agent | SWE-bench | — |
| ↳ | 76.8 | medium · mini-SWE-agent · 2 runs | SWE-bench | — |
| ↳ | 76.8 | high · mini-SWE-agent | SWE-bench | — |
| Settings differ by 2.4 points on this benchmark. | ||||
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 2 hosts
Other listings of this model
this page is one of them
The corpus files Claude 4.5 Opus under several keys. This page is the claude-4-5-opus listing. The full record — every host, every price and every benchmark score — is on the main Claude 4.5 Opus page.
More from Anthropic
most-hosted first