modelbenchmark.io

GPT-5.4 nano

OpenAI · gpt-5-4-nano

Compare

Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation

Specification

openai listing

Context
400K
Max output
128K
Released
2026-03-17
Knowledge cutoff
2025-08-31
Retires
2027-04-01
Open weights
no
Input
text, image
Output
text

Price

US dollars per million tokens · openai list price

Input
$0.20
Output
$1.25
Cache read
$0.02
Cache write
$0.20

Batch and priority prices

US dollars per million tokens · per listing

PriceListing$/M
Batch inputazure$0.10
Batch inputazure$0.11
Batch outputazure$0.625
Batch outputazure$0.6875
Priority inputazure_ai$0.40
Priority outputazure_ai$2.50

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

Quality

10 benchmarks · 2 sources

Composite
46th
percentile of 344 scored models
Rank
188 / 344
±0.07 sd
Evidence
10 × 2
3 or more benchmarks, from 2 or more sources
Effort range
18.9
points between effort settings
reasoning55th3/4 bench
math55th5/7 bench
knowledge4th1/1 bench
general20th1/1 bench

By reasoning effort

every setting placed on the same scale as the leaderboard

SettingCompositeBenchmarks behind it
xhigh66th2
high50th9
low40th5
medium18th1

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
FrontierMath-v142.9 ±2.6high · 2 runsEpoch AI2026-04-15
OTIS Mock AIME 2024-202587.8 ±3.9highEpoch AI2026-04-14
↳68.9 ±7.0lowEpoch AI2026-08-07
Settings differ by 18.9 points on this benchmark.
GPQA diamond78.5 ±2.4highEpoch AI2026-04-14
↳72.2 ±3.2lowEpoch AI2026-08-07
Settings differ by 6.3 points on this benchmark.
Chess Puzzles45.0 ±5.0xhighEpoch AI2026-10-14
↳36.0 ±4.8highEpoch AI2026-10-13
↳20.0 ±4.0lowEpoch AI2026-10-13
Settings differ by 25.0 points on this benchmark.
FrontierMath-Tier-4-2025-07-016.2 ±3.5highEpoch AI2026-04-15
FrontierMath-Tiers-1-3-v244.9 ±3.0highEpoch AI2026-06-12
↳20.4 ±2.4lowEpoch AI2026-08-28
Settings differ by 24.5 points on this benchmark.
LiveBench70.8xhigh · 23 runsLiveBench2026-06-25
FrontierMath-Tier-4-v212.2 ±5.2highEpoch AI2026-06-12
Mystery Game Puzzles6.0 ±2.4mediumEpoch AI2026-08-27
↳5.0 ±2.2highEpoch AI2026-08-27
↳3.0 ±1.7lowEpoch AI2026-08-27
Settings differ by 3.0 points on this benchmark.
SimpleQA Verified11.7 ±1.0highEpoch AI2026-08-27

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

Available from 39 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
openaivendor$0.20$1.25$0.02—
400K
272K
128K2027-04-01
poe$0.1818$1.1364$0.0182—400K128K—
302ai$0.20$1.25$0.02—400K128K—
abacus$0.20$1.25$0.02—400K128K—
aihubmix$0.20$1.25$0.02—400K128K—
aihubmix · aihubmix$0.20$1.25$0.02—400K128K—
azure
$0.20
$0.22
$1.25
$1.375
$0.02
$0.022
—
400K
272K
128K2027-09-21
azure_ai$0.20$1.25$0.02—272K128K2027-09-21
azure-cognitive-services$0.20$1.25$0.02—400K128K—
cloudflare-ai-gateway$0.20$1.25$0.02—128K128K—
crossmodel$0.20$1.25$0.02$0.20400K128K—
databricks$0.20$1.25$0.02$0.20
400K
272K
128K—
edenai$0.20$1.25$0.02—400K128K—
fastrouter$0.20$1.25$0.02—400K128K—
frogbot$0.20$1.25$0.02—400K128K—
github_copilot · github-copilot$0.20$1.25$0.02—272K128K—
github-copilot$0.20$1.25$0.02—400K128K—
impossibl$0.20$1.25$0.02—400K128K—
kilo$0.20$1.25$0.02—400K128K—
llmgateway$0.20$1.25$0.02—400K128K—
llmgateway-providers$0.20$1.25$0.02—400K128K—
merge-gateway$0.20$1.25$0.02—400K128K—
nano-gpt$0.20$1.25$0.02—400K128K—
nearai$0.20$1.25$0.02—400K128K—
neon$0.20$1.25$0.02—400K128K—
ofox$0.20$1.25$0.02—400K128K—
opencode$0.20$1.25$0.02—400K128K—
openrouter$0.20$1.25$0.02—400K128K—
opper$0.20$1.25$0.02—400K128K—
orcarouter$0.20$1.25$0.02—400K128K—
perplexity$0.20$1.25$0.02————
prime$0.20$1.25——400K128K—
requesty$0.20$1.25$0.02—400K128K—
vercel$0.20$1.25$0.02—400K128K—
vivgrid$0.20$1.25$0.02—400K128K—
zenmux$0.20$1.25$0.02—400K128K—
zerogpu · zerogpu$0.20$1.25$0.02—272K128K—
aimlapi$0.2116$1.3225$0.0212—400K128K—
model-oracle-ai————400K128K—

Price history

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

DateHostField$/M
2026-09-02azureCache read$0.02
2026-09-17azureCache read$0.02
2026-09-02ofoxCache read$0.02
2026-09-19ofoxCache read$0.016
2026-10-17ofoxCache read$0.02
2026-09-02openrouterCache read$0.02
2026-09-06openrouterCache read$0.02
2026-10-10openrouterCache read$0.02
2026-09-02poeCache read$0.018
2026-10-22poeCache read$0.0182
2026-09-02azureInput$0.20
2026-09-17azureInput$0.20
2026-09-02ofoxInput$0.20
2026-09-19ofoxInput$0.16
2026-10-17ofoxInput$0.20
2026-09-02openrouterInput$0.20
2026-09-06openrouterInput$0.20
2026-10-10openrouterInput$0.20
2026-09-02poeInput$0.18
2026-10-22poeInput$0.1818
2026-09-02azureOutput$1.25
2026-09-17azureOutput$1.25
2026-09-02ofoxOutput$1.25
2026-09-19ofoxOutput$1.00
2026-10-17ofoxOutput$1.25
2026-09-02openrouterOutput$1.25
2026-09-06openrouterOutput$1.25
2026-10-10openrouterOutput$1.25
2026-09-02poeOutput$1.10
2026-10-22poeOutput$1.1364
2026-10-13aimlapiCache read$0.0275
2026-10-23aimlapiCache read$0.0212
2026-10-13aimlapiInput$0.2751
2026-10-23aimlapiInput$0.2116
2026-10-13aimlapiOutput$1.7192
2026-10-23aimlapiOutput$1.3225

More from OpenAI

most-hosted first