modelbenchmark.io

Claude Haiku 5.5

Anthropic · claude-haiku-5-5

Compare

Fast Claude model for responsive assistance, classification, and lightweight agents

Specification

anthropic listing

Context
1M
Max output
128K
Released
2026-10-07
Knowledge cutoff
2026-06
Retires
2027-09-29
Open weights
no
Input
text, image, pdf
Output
text

Price

US dollars per million tokens · anthropic list price

Input
$0.10
Output
$0.50
Cache read
$0.01
Cache write
$0.125

Batch and priority prices

US dollars per million tokens · per listing

PriceListing$/M
Batch inputanthropic$0.05
Batch outputanthropic$0.25
Cache write, 1 houranthropic$0.20
Cache write, 1 hourbedrock_mantle · bedrock-mantle$0.22
Cache write, 1 hourbedrock_mantle · bedrock-mantle$0.24

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
100KCache read$0.05
100KCache write$0.625
100KCache write, 1 hour$1.00
100KInput$0.50
100KOutput$2.50

Quality

10 benchmarks · 2 sources

Composite
75th
percentile of 344 scored models
Rank
86 / 344
±0.07 sd
Evidence
10 × 2
3 or more benchmarks, from 2 or more sources
Effort range
4.5
points between effort settings
reasoning82nd4/4 bench
math88th3/7 bench
knowledge20th1/1 bench
general18th1/1 bench

By reasoning effort

every setting placed on the same scale as the leaderboard

SettingCompositeBenchmarks behind it
high95th1
xhigh84th2
max72nd9
low47th1

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
OTIS Mock AIME 2024-202598.9 ±0.5xhighEpoch AI2026-10-07
↳97.2 ±1.3highEpoch AI2026-10-07
Settings differ by 1.7 points on this benchmark.
GPQA diamond89.7 ±2.0maxEpoch AI2026-10-09
FrontierMath-Tiers-1-3-v275.1 ±2.6maxEpoch AI2026-10-07
EBR-bench43.8 ±1.8maxEpoch AI2026-10-09
Chess Puzzles55.0 ±5.0maxEpoch AI2026-10-13
↳24.0 ±4.3lowEpoch AI2026-10-13
Settings differ by 31.0 points on this benchmark.
Mystery Game Puzzles30.0 ±4.6maxEpoch AI2026-10-07
FrontierMath-Tier-4-v246.3 ±7.9maxEpoch AI2026-10-07
Furniture Assembly47.5 ±6.4maxEpoch AI2026-10-07
LiveBench72.9xhigh · 23 runsLiveBench2026-06-25
↳68.4max · 23 runsLiveBench2026-06-25
Settings differ by 4.5 points on this benchmark.
SimpleQA Verified23.8 ±1.3maxEpoch AI2026-10-09

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

Available from 53 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
anthropicvendor$0.10$0.50$0.01$0.1251M128K—
302ai$0.10$0.50$0.01$0.1251M128K—
aihubmix$0.10$0.50$0.01$0.1251M128K—
aimlapi$0.10$0.50$0.01$0.1251M128K—
aixy$0.10$0.50$0.01$0.1251M128K—
amazon-bedrock$0.10$0.50$0.01$0.1251M128K—
amazon-bedrock · global$0.10$0.50$0.01$0.1251M128K—
azure$0.10$0.50$0.01$0.1251M128K—
azure_ai$0.10$0.50$0.01$0.1251M128K2027-09-29
azure-cognitive-services$0.10$0.50$0.01$0.1251M128K—
bedrock_converse$0.10$0.50$0.01$0.1251M128K—
bedrock_converse · global$0.10$0.50$0.01$0.1251M128K—
cloudflare-ai-gateway$0.10$0.50$0.01$0.1251M128K—
commandcode$0.10$0.50$0.01$0.1251M128K—
crossmodel$0.10$0.50$0.01$0.1251M128K—
databricks$0.10$0.50$0.01$0.1251M128K—
edenai$0.10$0.50$0.01$0.1251M128K—
github_copilot · github-copilot$0.10$0.50$0.01$0.125128K16K—
github-copilot$0.10$0.50$0.01$0.1251M128K—
kilo$0.10$0.50$0.01$0.1251M128K—
llmgateway$0.10$0.50$0.01$0.1251M128K—
llmgateway-providers$0.10$0.50$0.01$0.1251M128K—
llmgateway-providers · aws-bedrock$0.10$0.50$0.01$0.1251M128K—
merge-gateway$0.10$0.50$0.01$0.1251M128K—
nano-gpt$0.10$0.50$0.01$0.1251M128K—
nearai$0.10$0.50$0.01$0.1251M128K—
neon$0.10$0.50$0.01$0.1251M128K—
ofox$0.10$0.50$0.01$0.1251M128K—
opencode$0.10$0.50$0.01$0.1251M128K—
opencode_go · opencode-go$0.10$0.50$0.01$0.1251M128K—
opencode-go$0.10$0.50$0.01$0.1251M128K—
openrouter$0.10$0.50$0.01$0.1251M128K—
perplexity$0.10$0.50$0.01————
perplexity-agent$0.10$0.50$0.01$0.1251M128K—
requesty$0.10$0.50$0.01$0.131M128K—
standardcompute$0.10$0.50$0.01$0.1251M50K—
venice$0.10$0.50$0.01$0.1251M128K—
vercel$0.10$0.50$0.01$0.1251M128K—
vertex_ai-anthropic_models$0.10$0.50$0.01$0.1251M128K—
vivgrid$0.10$0.50$0.01$0.1251M128K—
amazon-bedrock · au$0.11$0.55$0.011$0.13751M128K—
amazon-bedrock · eu$0.11$0.55$0.011$0.13751M128K—
amazon-bedrock · jp$0.11$0.55$0.011$0.13751M128K—
amazon-bedrock · us$0.11$0.55$0.011$0.13751M128K—
bedrock_converse · apac$0.11$0.55$0.011$0.13751M128K—
bedrock_converse · au$0.11$0.55$0.011$0.13751M128K—
bedrock_converse · eu$0.11$0.55$0.011$0.13751M128K—
bedrock_converse · jp$0.11$0.55$0.011$0.13751M128K—
bedrock_converse · us$0.11$0.55$0.011$0.13751M128K—
bedrock_mantle · bedrock-mantle
$0.11
$0.12
$0.55
$0.60
$0.011
$0.012
$0.1375
$0.15
1M128K—
cortecs$0.113$0.563$0.011$0.141M128K—
bedrock
$0.12
$0.10
$0.60
$0.50
$0.012
$0.01
$0.15
$0.125
1M128K—
kyma · kyma$0.135$0.675$0.0135—1M128K—

Price history

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

DateHostField$/M
2026-10-08bedrockCache read$0.012
2026-10-25bedrockCache read$0.012
2026-10-08openrouterCache read$0.01
2026-10-10openrouterCache read$0.01
2026-10-08veniceCache read$0.0125
2026-10-15veniceCache read$0.01
2026-10-08bedrockCache write$0.15
2026-10-25bedrockCache write$0.15
2026-10-08openrouterCache write$0.125
2026-10-10openrouterCache write$0.125
2026-10-08veniceCache write$0.1562
2026-10-15veniceCache write$0.125
2026-10-08bedrockInput$0.12
2026-10-25bedrockInput$0.12
2026-10-08openrouterInput$0.10
2026-10-10openrouterInput$0.10
2026-10-08veniceInput$0.125
2026-10-15veniceInput$0.10
2026-10-08bedrockOutput$0.60
2026-10-25bedrockOutput$0.60
2026-10-08openrouterOutput$0.50
2026-10-10openrouterOutput$0.50
2026-10-08veniceOutput$0.625
2026-10-15veniceOutput$0.50
2026-10-13aimlapiCache read$0.013
2026-10-23aimlapiCache read$0.01
2026-10-13aimlapiInput$0.13
2026-10-23aimlapiInput$0.10
2026-10-13aimlapiOutput$0.65
2026-10-23aimlapiOutput$0.50
2026-10-21aimlapiCache write$0.1625
2026-10-23aimlapiCache write$0.125

Other listings of this model

same model, different key

More from Anthropic

most-hosted first