Pass IndexThe State of AISign in

o3-pro

The o-series of models are trained with reinforcement learning to think before they answer and perform complex reasoning.

text + image + file → text · made by OpenAI

$20$80per Mtok in / out
OpenAI · 7 sellers

Sold by 10 ways

SellerLaneRate
Microsoft Azure AI cloudbatch$10per Mtok in$40per Mtok outraw.githubusercontent.com · read 2026-09-16
Microsoft Azure AI cloudstandard$20per Mtok in$80per Mtok outraw.githubusercontent.com · read 2026-09-16
OpenAI apistandard$20per Mtok in$80per Mtok outdevelopers.openai.com · read 2026-08-24
Nous Research aggregatornot on the seller's list todaybatch$8per Mtok in$32per Mtok outinference-api.nousresearch.com · read 2026-08-25
Nous Research aggregatorstandard$20per Mtok in$80per Mtok outinference-api.nousresearch.com · read 2026-09-12
OpenRouter aggregatornot on the seller's list todaybatch$10per Mtok in$40per Mtok outopenrouter.ai · read 2026-08-24
OpenRouter aggregatorstandard$20per Mtok in$80per Mtok outopenrouter.ai · read 2026-08-24
ElectronHub aggregatorstandard$20per Mtok in$80per Mtok outapi.electronhub.ai · read 2026-09-16
Requesty aggregatoropenai-responses$20per Mtok in$80per Mtok out$20per Mtok cachedrouter.requesty.ai · read 2026-08-28
Vercel AI Gateway aggregatorstandard$20per Mtok in$80per Mtok outai-gateway.vercel.sh · read 2026-08-25

Measured 15 standings

PlaceBoardMetricScore
3rdof 69Aider polyglot coding leaderboardPercent correct (high)84.9
3rdof 72Aider polyglot — Epoch AIPercent correct (high)84.9
5thof 42Fictionlivebench — Epoch AI120k token score (medium)0.889
5thof 49Lech mazur writing — Epoch AIMean score (medium)8.44
40thof 123Lmca — Epoch AIScore (high)38.48
43rdof 161Dtbench — Epoch AIAccuracy (high)86.93
51stof 170Weirdml — Epoch AIAccuracy (high)58.21
103rdof 105Vectara Hallucination LeaderboardHallucination Rate23.3
115thof 205ARC-AGI-1 (Public Eval)Score (%) (high)63.34
119thof 230ARC-AGI-1 (Semi-Private)Score (%) (high)59.33
130thof 244Arc agi — Epoch AIScore (high)59.33
133rdof 216ARC-AGI-2 (Public Eval)Score (%) (high)3.89
138thof 232ARC-AGI-2 (Semi-Private)Score (%) (high)4.86
142ndof 227Arc agi 2 — Epoch AIScore (high)4.86
262ndof 656Epoch capabilities index — Epoch AIECI Score (high)147.46

About

o3-pro — a text model from OpenAI, sold by 7 companies from $20 in and $80 out per million tokens, placed 3rd of 72 on Aider polyglot — Epoch AI.

It takes text, images and files and returns text, with a context window of 200,000 tokens. It was published in June 2024, trained on material up to May 2024. Its sellers say it can reason step by step and call a tool. The catalogue files it under extract, reasoning and code. Seven companies sell it. The cheapest is $20 in and $80 out per million tokens at ElectronHub. Beside the standard rate there are batch and separately routed lanes. It has been measured on 15 boards, and stands best at 3rd of 72 on Aider polyglot — Epoch AI.

Every current figure

Maker
OpenAI
Register
model
Takes
text + image + file
Returns
text
Context
200,000 tokens
Longest answer
100,000 tokens
Published
June 2024
Knowledge to
May 2024
Licence
not read
Sellers
7
Price
$20 in and $80 out per million tokens — ElectronHub
Boards
15
Best place
3rd of 72 — Aider polyglot — Epoch AI

Known as 5 names

o3-pro (high)openai/o3-proopenai/o3-pro:batchopenai-responses/o3-proOpenAI: o3 Pro (batch)