modelbenchmark.io

o3-mini

OpenAI · o3-mini

Compare

Smaller o-series reasoner for economical coding, math, and planning tasks

Specification

openai listing

Context
200K
Max output
100K
Released
2024-12-20
Knowledge cutoff
2024-05
Retires
2026-10-23
Open weights
no
Input
text
Output
text

Price

US dollars per million tokens · openai list price

Input
$1.10
Output
$4.40
Cache read
$0.55
Cache write
free

Batch and priority prices

US dollars per million tokens · per listing

PriceListing$/M
Batch inputazure$0.605
Batch inputopenai$0.55
Batch outputazure$2.42
Batch outputopenai$2.20

Batch input is 45% below the interactive input price at the same listing.

Quality

12 benchmarks · 3 sources

Composite
42nd
percentile of 334 scored models
Rank
193 / 334
±0.05 sd
Evidence
12 × 3
3 or more benchmarks, from 2 or more sources
Effort range
6.6
points between effort settings
coding36th2/3 bench
reasoning46th3/4 bench
math43rd6/7 bench
knowledge11th1/1 bench

By reasoning effort

every setting placed on the same scale as the leaderboard

SettingCompositeBenchmarks behind it
high82nd1
medium71st1

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
MATH level 595.8 ±0.4default · 2 runsEpoch AI2025-02-13
Aider polyglot60.4high · aider diffAider2025-01-31
53.8medium · aider diffAider2025-01-31
Settings differ by 6.6 points on this benchmark.
GPQA diamond73.2 ±3.3default · 3 runsEpoch AI2026-07-20
OTIS Mock AIME 2024-202561.8 ±7.5default · 3 runsEpoch AI2026-07-20
FrontierMath-v117.9 ±1.9default · 4 runsEpoch AI2025-11-16
Chess Puzzles10.7 ±2.9default · 3 runsEpoch AI2026-08-07
FrontierMath-Tier-4-2025-07-012.1 ±2.9default · 2 runsEpoch AI2025-07-01
SWE-Bench verified42.4default · Agentless LiteSWE-bench
Mystery Game Puzzles7.0 ±2.6defaultEpoch AI2026-08-27
FrontierMath-Tier-4-v20.0 ±0.0defaultEpoch AI2026-06-11
FrontierMath-Tiers-1-3-v211.0 ±1.8default · 3 runsEpoch AI2026-08-27
SimpleQA Verified15.3 ±1.1defaultEpoch AI2026-08-31

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

Available from 20 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
openaivendor$1.10$4.40$0.55200K100K2026-10-23
poe$0.99$4.00200K100K
abacus$1.10$4.40$0.55200K100K
azure
$1.10
$1.21
$1.513
$4.40
$4.84
$6.05
$0.55
$0.605
$0.757
200K100K2026-11-19
azure-cognitive-services$1.10$4.40$0.55200K100K
cloudflare-ai-gateway$1.10$4.40$0.55200K100K
edenai$1.10$4.40$0.55200K100K
helicone$1.10$4.40$0.55200K100K
impossibl$1.10$4.40$0.55200K100K
jiekou$1.10$4.40131K131K
kilo$1.10$4.40$0.55200K100K
llmgateway$1.10$4.40$0.55200K100K
llmgateway-providers$1.10$4.40$0.55200K100K
merge-gateway$1.10$4.40$0.55200K100K
nano-gpt$1.10$4.40$0.55200K100K
nearai$1.10$4.40$0.55200K100K
openrouter$1.10$4.40$0.55200K100K
vercel$1.10$4.40$0.55200K100K
vercel_ai_gateway · vercel-ai-gateway$1.10$4.40$0.55free200K100K
anyapi200K100K

Price history

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

DateHostField$/M
2026-09-02azureCache read$0.55
2026-09-04azureCache read$0.55
2026-09-02openrouterCache read$0.55
2026-09-06openrouterCache read$0.55
2026-09-02azureInput$1.10
2026-09-04azureInput$1.10
2026-09-02azureOutput$4.40
2026-09-04azureOutput$4.40

Other listings of this model

same model, different key

More from OpenAI

most-hosted first