modelbenchmark.io

Claude Sonnet 4.6

Anthropic · claude-4-6-sonnet

Compare

Balanced Claude model for coding, analysis, agent workflows, and cost control

Specification

most-agreed values

Context
1M
Max output
128K
Released
2026-02-17
Knowledge cutoff
2025-08-31
Retires
Open weights
no
Input
text, image, pdf
Output
text

Price

US dollars per million tokens · most-agreed

Input
$3.196
Output
$15.94
Cache read
$0.32
Cache write
$3.999

Quality

8 benchmarks · 2 sources

Composite
64th
percentile of 334 scored models
Rank
121 / 334
±0.08 sd
Evidence
8 × 2
3 or more benchmarks, from 2 or more sources
Effort range
9.3
points between effort settings
reasoning59th3/4 bench
math84th3/7 bench
knowledge38th1/1 bench
general32nd1/1 bench

By reasoning effort

every setting placed on the same scale as the leaderboard

SettingCompositeBenchmarks behind it
32k83rd3
16k83rd2
medium75th3
high63rd4
max57th4
auto42nd1
low42nd1

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
FrontierMath-v156.2 ±13.316k · 2 runsEpoch AI2026-02-22
GPQA diamond87.4 ±2.032kEpoch AI2026-02-20
83.3 ±2.7highEpoch AI2026-07-13
83.3 ±2.7mediumEpoch AI2026-07-13
78.8 ±2.9maxEpoch AI2026-08-06
Settings differ by 8.6 points on this benchmark.
OTIS Mock AIME 2024-202585.8 ±4.532kEpoch AI2026-02-20
82.2 ±5.8mediumEpoch AI2026-07-13
75.6 ±6.5highEpoch AI2026-07-13
71.1 ±6.8maxEpoch AI2026-08-06
Settings differ by 14.7 points on this benchmark.
FrontierMath-Tier-4-2025-07-014.2 ±0.016k · 2 runsEpoch AI2026-02-22
LiveBench73.4auto · 23 runsLiveBench2026-06-25
Mystery Game Puzzles16.0 ±3.7lowEpoch AI2026-08-06
SimpleQA Verified35.5 ±1.5highEpoch AI2026-08-10
32.8 ±1.5maxEpoch AI2026-08-27
Settings differ by 2.7 points on this benchmark.
Chess Puzzles13.0 ±3.432kEpoch AI2026-02-20
8.0 ±2.7mediumEpoch AI2026-07-13
5.0 ±2.2highEpoch AI2026-07-13
3.0 ±1.7maxEpoch AI2026-08-06
Settings differ by 10.0 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 1 host

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
cortecs$3.196$15.94$0.32$3.9991M128K

Other listings of this model

this page is one of them

The corpus files Claude Sonnet 4.6 under several keys. This page is the claude-4-6-sonnet listing. The full record — every host, every price and every benchmark score — is on the main Claude Sonnet 4.6 page.

More from Anthropic

most-hosted first