modelbenchmark.io

Claude Sonnet 4.5

Anthropic · claude-sonnet-4-5

Compare

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

Specification

anthropic listing

Context
1M
Max output
64K
Released
2025-09-29
Knowledge cutoff
2025-07-31
Retires
2026-09-29
Open weights
no
Input
text, image, pdf
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

Batch and priority prices

US dollars per million tokens · per listing

PriceListing$/M
Batch inputbedrock_converse$1.50
Batch inputbedrock_converse · au$1.65
Batch outputbedrock_converse$7.50
Batch outputbedrock_converse · au$8.25
Cache write, 1 hourbedrock_converse$6.00
Cache write, 1 hourbedrock_converse · au$6.60
Cache write, 1 hourbedrock$7.20

Batch input is 50% below the interactive input price at the same listing.

Long-context price tiers

the rate above each context threshold

AbovePrice$/M
200KCache read$0.60
200KCache write$7.50
200KCache write, 1 hour$12.00
200KInput$6.00
200KOutput$22.50

Quality

10 benchmarks · 1 source

Composite
54th
percentile of 334 scored models
Rank
155 / 334
±0.05 sd
Evidence
10 × 1
one source, or fewer than 3 benchmarks
Effort range
3.5
points between effort settings
reasoning66th3/4 bench
math52nd6/7 bench
knowledge27th1/1 bench

By reasoning effort

every setting placed on the same scale as the leaderboard

SettingCompositeBenchmarks behind it
16k82nd2
59k63rd4
32k59th8
48k45th1

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
MATH level 597.7 ±0.432kEpoch AI2025-10-21
GPQA diamond82.3 ±2.759kEpoch AI2025-10-28
81.7 ±2.832kEpoch AI2025-10-21
78.8 ±2.916kEpoch AI2025-10-28
Settings differ by 3.5 points on this benchmark.
OTIS Mock AIME 2024-202577.8 ±6.332kEpoch AI2025-10-21
77.8 ±6.359kEpoch AI2025-10-28
71.1 ±6.816kEpoch AI2025-10-28
Settings differ by 6.7 points on this benchmark.
Chess Puzzles12.0 ±3.332kEpoch AI2025-12-08
Mystery Game Puzzles17.0 ±3.848kEpoch AI2026-07-25
FrontierMath-Tier-4-2025-07-014.2 ±2.932kEpoch AI2025-10-22
FrontierMath-v115.2 ±2.132kEpoch AI2025-11-16
13.5 ±2.059kEpoch AI2025-11-13
Settings differ by 1.7 points on this benchmark.
SimpleQA Verified30.7 ±1.559kEpoch AI2026-08-27
FrontierMath-Tiers-1-3-v223.9 ±2.532kEpoch AI2026-06-11
FrontierMath-Tier-4-v22.4 ±2.432kEpoch AI2026-06-11

The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.

Available from 57 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
anthropicvendor$3.00$15.00$0.30$3.75
1M
200K
64K2026-09-29
qihang-ai$0.43$2.14200K64K
poe$2.60$13.00$0.26$3.20983K33K
jiekou$2.70$13.50200K64K
302ai$3.00$15.00200K64K
abacus$3.00$15.00200K64K
amazon-bedrock$3.00$15.00$0.30$3.75200K64K
amazon-bedrock · au$3.00$15.00$0.30$3.75200K64K
amazon-bedrock · global$3.00$15.00$0.30$3.75200K64K
amazon-bedrock · jp$3.00$15.00$0.30$3.75200K64K
amazon-bedrock · us$3.00$15.00$0.30$3.75200K64K
azure$3.00$15.00$0.30$3.75200K64K
azure_ai$3.00$15.00$0.30$3.75200K64K2026-10-19
azure-cognitive-services$3.00$15.00$0.30$3.75200K64K
bedrock_converse$3.00$15.00$0.30$3.75200K64K
bedrock_converse · global$3.00$15.00$0.30$3.75200K64K
cloudflare-ai-gateway$3.00$15.00$0.30$3.751M64K
databricks$3.00$15.00$0.30$3.75200K64K
gmi · gmi$3.00$15.00410K32K
google-vertex$3.00$15.00$0.30$3.75200K64K
google-vertex-anthropic$3.00$15.00$0.30$3.75200K64K
helicone$3.00$15.00$0.30$3.75200K64K
impossibl$3.00$15.00$0.30$3.75200K64K
kilo$3.00$15.00$0.30$3.751M64K
llmgateway$3.00$15.00$0.30$3.75200K64K
llmgateway-providers$3.00$15.00$0.30$3.75200K64K
llmgateway-providers · aws-bedrock$3.00$15.00$0.30$3.75200K8K
llmgateway-providers · vertex-anthropic$3.00$15.00$0.30200K64K
merge-gateway$3.00$15.00$0.30$3.75200K64K
nano-gpt$3.00$15.00$0.30200K64K
nano-gpt · thinking$3.00$15.00$0.30200K64K
nearai$3.00$15.00$0.30$3.75200K64K
neon$3.00$15.00$0.30$3.75200K64K
ofox$3.00$15.00$0.30$3.75200K64K
opencode$3.00$15.00$0.30$3.751M64K
openrouter$3.00$15.00$0.30$3.75
1M
200K
64K
opper$3.00$15.00$0.30$3.75200K64K
orcarouter$3.00$15.00$0.30$3.751M64K
perplexity$3.00$15.00$0.30
perplexity-agent$3.00$15.00$0.30200K64K
pioneer$3.00$15.00$0.30$3.751M64K
requesty$3.00$15.00$0.30$3.751M64K
snowflake$3.00$15.00$0.30200K16K
vercel$3.00$15.00$0.30$3.751M64K
vercel_ai_gateway · vercel-ai-gateway$3.00$15.00$0.30$3.751M64K
vertex_ai-anthropic_models$3.00$15.00$0.30$3.75200K64K2026-09-29
zenmux$3.00$15.00$0.30$3.751M64K
amazon-bedrock · eu$3.30$16.50$0.33$4.125200K64K
bedrock_converse · au$3.30$16.50$0.33$4.125200K64K
bedrock_converse · eu$3.30$16.50$0.33$4.125200K64K
bedrock_converse · jp$3.30$16.50$0.33$4.125200K64K
bedrock_converse · us$3.30$16.50$0.33$4.125200K64K
bedrock$3.60$18.00$0.36$4.50200K8K
venice$3.75$18.75$0.375$4.69198K64K
anyapi200K64K
github_copilot · github-copilot128K16K
snowflake-cortex200K16K

Price history

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

DateHostField$/M
2026-09-02nearaiOutput$15.50
2026-09-10nearaiOutput$15.00

More from Anthropic

most-hosted first