modelbenchmark.io

Claude Opus 5.5

Anthropic · claude-opus-5-5

Compare

Claude model for long-running agentic coding and knowledge work

Specification

anthropic listing

Context
1M
Max output
128K
Released
2026-09-22
Knowledge cutoff
2026-06
Retires
—
Open weights
no
Input
text, image, pdf
Output
text

Price

US dollars per million tokens · anthropic list price

Input
$4.00
Output
$20.00
Cache read
$0.20
Cache write
$5.00

Batch and priority prices

US dollars per million tokens · per listing

PriceListing$/M
Batch inputanthropic$2.00
Batch inputvertex_ai-anthropic_models$2.50
Batch outputanthropic$10.00
Batch outputvertex_ai-anthropic_models$12.50
Cache write, 1 hourbedrock_converse$8.00
Cache write, 1 hourbedrock_converse · us$8.80
Cache write, 1 hourbedrock$9.60
Cache write, 1 houropenrouter · batch$4.00

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

Quality

10 benchmarks · 2 sources

Composite
99th
percentile of 340 scored models
Rank
4 / 340
±0.08 sd
Evidence
10 × 2
3 or more benchmarks, from 2 or more sources
Effort range
1.1
points between effort settings
coding99th1/3 bench
reasoning99th3/4 bench
math98th3/7 bench
knowledge97th1/1 bench
general95th1/1 bench

By reasoning effort

every setting placed on the same scale as the leaderboard

SettingCompositeBenchmarks behind it
max99th10
xhigh95th1

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
Mystery Game Puzzles71.0 ±4.6maxEpoch AI2026-09-29
Furniture Assembly83.3 ±4.7maxEpoch AI2026-09-22
FrontierMath-Tier-4-v295.0 ±4.1maxEpoch AI2026-09-22
EBR-bench71.4 ±7.1maxEpoch AI2026-09-22
FrontierMath-Tiers-1-3-v291.2 ±1.7maxEpoch AI2026-09-22
SimpleQA Verified72.2 ±1.4maxEpoch AI2026-09-22
MirrorCode77.4 ±8.3maxEpoch AI2026-09-22
LiveBench83.2max · 23 runsLiveBench2026-06-25
↳82.1xhigh · 23 runsLiveBench2026-06-25
Settings differ by 1.1 points on this benchmark.
OTIS Mock AIME 2024-2025100.0 ±0.0maxEpoch AI2026-09-22
GPQA diamond90.6 ±1.9maxEpoch AI2026-09-22

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

Available from 42 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
anthropicvendor$4.00$20.00$0.20$5.001M128K—
openrouter · batch$2.00$10.00$0.10$2.501M128K—
302ai$4.00$20.00$0.20$5.001M128K—
aihubmix$4.00$20.00$0.20$5.001M128K—
amazon-bedrock$4.00$20.00$0.20$5.001M128K—
amazon-bedrock · global$4.00$20.00$0.20$5.001M128K—
azure$4.00$20.00$0.20$5.001M128K—
azure_ai$4.00$20.00$0.20$5.001M128K—
azure-cognitive-services$4.00$20.00$0.20$5.001M128K—
bedrock_converse$4.00$20.00$0.20$5.001M128K—
bedrock_converse · global$4.00$20.00$0.20$5.001M128K—
cloudflare-ai-gateway$4.00$20.00$0.20$5.001M128K—
crossmodel$4.00$20.00$0.20$5.001M128K—
databricks$4.00$20.00$0.20$5.001M128K—
edenai$4.00$20.00$0.20$5.001M128K—
github-copilot$4.00$20.00$0.20$5.001M128K—
kilo$4.00$20.00$0.20$5.001M128K—
llmgateway$4.00$20.00$0.20$5.001M128K—
llmgateway-providers$4.00$20.00$0.20$5.001M128K—
merge-gateway$4.00$20.00$0.20$5.001M128K—
nano-gpt$4.00$20.00$0.20$5.001M128K—
ofox$4.00$20.00$0.20$5.001M128K—
opencode$4.00$20.00$0.20$5.001M128K—
openrouter$4.00$20.00$0.20$5.001M128K—
requesty$4.00$20.00$0.20$5.001M128K—
vercel$4.00$20.00$0.20$5.001M128K—
vertex_ai-anthropic_models$4.00$20.00$0.20$5.001M128K—
vivgrid$4.00$20.00$0.20$5.001M128K—
zenmux$4.00$20.00$0.20$5.001M128K—
cortecs$4.399$21.998$0.219$5.501M128K—
amazon-bedrock · au$4.40$22.00$0.22$5.501M128K—
amazon-bedrock · eu$4.40$22.00$0.22$5.501M128K—
amazon-bedrock · jp$4.40$22.00$0.22$5.501M128K—
amazon-bedrock · us$4.40$22.00$0.22$5.501M128K—
bedrock_converse · apac$4.40$22.00$0.22$5.501M128K—
bedrock_converse · au$4.40$22.00$0.22$5.501M128K—
bedrock_converse · eu$4.40$22.00$0.22$5.501M128K—
bedrock_converse · jp$4.40$22.00$0.22$5.501M128K—
bedrock_converse · us$4.40$22.00$0.22$5.501M128K—
bedrock_mantle · bedrock-mantle
$4.40
$4.80
$22.00
$24.00
$0.22
$0.24
$5.50
$6.00
1M128K—
bedrock$4.80$24.00$0.24$6.001M128K—
venice$4.80$24.00$0.24$6.001M128K—

Price history

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

DateHostField$/M
2026-09-23bedrock_converseCache read$0.20
2026-09-24bedrock_converseCache read$0.22
2026-09-25bedrock_converseCache read$0.20
2026-09-23requestyCache read$0.50
2026-09-26requestyCache read$0.20
2026-09-23bedrock_converseCache write$5.00
2026-09-24bedrock_converseCache write$5.50
2026-09-25bedrock_converseCache write$5.00
2026-09-23requestyCache write$6.25
2026-09-26requestyCache write$5.00
2026-09-23bedrock_converseInput$4.00
2026-09-24bedrock_converseInput$4.40
2026-09-25bedrock_converseInput$4.00
2026-09-23requestyInput$5.00
2026-09-26requestyInput$4.00
2026-09-23bedrock_converseOutput$20.00
2026-09-24bedrock_converseOutput$22.00
2026-09-25bedrock_converseOutput$20.00
2026-09-23requestyOutput$25.00
2026-09-26requestyOutput$20.00

Other listings of this model

same model, different key

More from Anthropic

most-hosted first