modelbenchmark.io

Claude 3.5 Sonnet

Anthropic · claude-3-5-sonnet

Compare

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

Specification

most-agreed values

Context
200K
Max output
8K
Released
2024-10-22
Knowledge cutoff
2024-04-30
Retires
2026-02-19
Open weights
no
Input
text, image
Output
text

Price

US dollars per million tokens · most-agreed

Input
$3.00
Output
$15.00
Cache read
$0.30
Cache write
$3.75

Quality

2 benchmarks · 2 sources

Composite
55th
percentile of 344 scored models
Rank
156 / 344
±0.09 sd
Evidence
2 × 2
one source, or fewer than 3 benchmarks
Effort range
—
one configuration only
coding41st2/3 bench

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
Aider polyglot51.6default · aider diffAider2025-01-17
SWE-Bench verified50.7default · SWE-agent · 13 runsSWE-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 9 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
bedrock$3.00$15.00$0.30$3.75200K
4K
8K
—
bedrock · eu$3.00$15.00$0.30$3.75200K
4K
8K
—
bedrock · us$3.00$15.00$0.30$3.75200K
4K
8K
—
snowflake$3.00$15.00$0.30—200K16K—
vercel_ai_gateway · vercel-ai-gateway$3.00$15.00$0.30$3.75200K8K—
vertex_ai-anthropic_models$3.00$15.00——200K8K2026-02-19
replicate · replicate$3.75$18.75—————
heroku · heroku————200K8K—
qiniu-ai————200K8K—

Price history

append-only observations · a listing writes a row only when its price moves

DateHostField$/M
2026-09-02bedrockCache read$0.30
2026-09-23bedrockCache read$0.30
2026-09-02bedrockCache write$3.75
2026-09-23bedrockCache write$3.75
2026-09-02bedrockInput$3.00
2026-09-23bedrockInput$3.00
2026-09-02bedrockOutput$15.00
2026-09-23bedrockOutput$15.00

More from Anthropic

most-hosted first