Deliberate o-series reasoner for hard math, coding, and multi-step analysis
o3 by openai is offered by 22 providers on this page. Public prices are shown for 21 of them. Its official list price is $2.00 per 1M input tokens and $8.00 per 1M output tokens. The lowest paid channel is Poe at $1.80 / $7.20 per 1M, about 5.6× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it.
The context window is 200,000 tokens at the reference host, but hosts report different limits, from 131,072 to 200,000, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text, Image, and PDF. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use. Its training knowledge cuts off at 2024-05.
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| Poe | Gateway | $1.80 | $7.20 | $0.45 | — | 200,000 | 100,000 | |
| Impossibl | Gateway | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| Azure Cognitive Services | First-party | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| Vercel AI Gateway | Cloud | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| DevPass (LLM Gateway) | Gateway | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| Cloudflare AI Gateway | Cloud | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| Kilo Gateway | Gateway | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| Azure | First-party | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| Merge Gateway | Gateway | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| Abacus | Gateway | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
| OpenAIOfficial | First-party | $2.00 | $8.00 | $0.50 | — | 200,000 | 100,000 | |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| Aider Polyglot | — | 81.3 | percent correct | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.
2 / 22 providers expose no reasoning control (reasoning_options: []).