o3
OpenAI · released Apr 16, 2025
- Type
- Proprietary
- Input
- $2.00
- Output
- $8.00
- Cached
- $0.50
List price · per 1M tokens · OpenAI at 200K context · source ↗
Our take
Written Aug 3, 2026o3 is a hosted-only reasoning model from OpenAI that handles text, images and files up to 200,000 tokens in a single request. It scores strongly on coding benchmarks but shows a clear gap between its specialist performance and its creative writing scores.
Choose this for complex coding tasks where measured benchmark scores matter, or for long-document and multi-file work that needs the full 200,000-token reach. It suits instruction-following workflows and situations where identical pricing across both tracked providers removes comparison friction. Skip it if you need downloadable weights, if creative writing quality is central, or if you want the cheapest output rates among reasoning models.
The case for it
- Strong measured coding performance: 84.7% on LiveCodeBench and an Arena Coding Elo of about 1459.
- 200,000-token request limit supports long-document and multi-file tasks.
- Identical pricing on both tracked providers, so provider choice does not create cost surprises.
- Known throughput of 71 tokens per second on OpenAI's own infrastructure.
The case against it
- Creative writing is its weakest arena score, 76 points below its own coding mark on the same leaderboard batch.
- Overall text Elo sits below both its coding and maths scores, suggesting general chat is not its strength.
- Output costs four times the input rate, a steep multiplier for a proprietary-only model.
How good is it?
IntelligencePuzzles, maths, exam questions
Arena Text (overall)54th of 143 · 1430.9
CodingWriting and fixing code on its own
Arena Coding65th of 143 · 1459.4
AgenticPlanning, calling tools, staying on task
o3 is not on Arena Agent (IPS), which is where the rating would come from, so there is no rating here. It is on SWE-bench Verified, in 22nd of 39 with 58.4.
WritingWe do not rate this
Two boards come close and neither tests writing: Arena Creative Writing asks people which of two replies they prefer, and LiveBench Language tests whether a model understood a passage. So we show where o3 placed and give it no mark out of five.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done, which is why they get no rating.
Every published score for this model8 scoresEvery figure we hold, from 8 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Or rent it from someone else
Why this differs from the header. The strip above quotes OpenAI's own list price. This is the cheapest live offer at the widest standard context we hold, whoever is serving it — a reseller undercutting a lab is ordinary commerce, not an error.
- per 1M tokens
- $2.00 in / $8.00 out
- Context served
- 200K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouter | $2.00 / $8.00 | 200K | not measured | Unknown | Unknown | Unknown |
| OpenAI | $2.00 / $8.00 | 200K100K out | 71 tok/s | No | Yesunknown period | Unknown |
Across the 2 listings we hold: 1 say they do not train on prompts, 0 say they do and 1 do not say. 0 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
- ✓
- Supported
- ✗
- Not supported
- Not published
- host gave no parameter list
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouter | ✓ | ✓ | ✓ |
| OpenAI | ✓ | ✓ | ✓ |
Tool calling: 2 of 2 listings say yes. JSON output: 2 of 2 listings say yes. Strict schema: 2 of 2 listings say yes.
Models people weigh against o3
When we formed this view
Dates behind this page
Prices last checked 9d ago
What we do not know about this model yet
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 2 listings do not say whether they train on prompts.
Licence and identifiers
What the licence allowsWe hold no licence record for this model. Inside are the identifiers you need to pull it — its Hugging Face repo where we have one, our slug and a machine-readable card.
Licence
Commercial API terms. We hold no licence record for this model, so there is nothing to summarise here.
Identifiers
- Modality record
- text+image+file->text
- Catalogue slug
- openai-o3