o4
openai/o4OpenAI's deliberate reasoning flagship, the ceiling for hard math and science problems.
Model overview
Context window
300K
tokens
Input price
—
quoted on request
Output price
—
quoted on request
Weekly volume
—
tokens / week
Open weights
No
API only
Capabilities
4
of 11 tags
Description
o4 is a reasoning model from OpenAI. Webparam does not stock this model yet and can source it on request — tell us what you need it for and we will come back with availability and a price. OpenAI’s stated focus is the reference point for frontier capability. The weights are not published, so it is API-only.
Capabilities
- Reasoning
- Spends extra billed tokens thinking before it answers, trading latency and cost for accuracy.
- Tool calling
- Calls functions you define and returns their arguments as structured data, the basis of agents.
- JSON mode
- Constrains the response to valid JSON matching a schema you supply.
- Streaming
- Emits tokens as they are generated, so answers appear progressively rather than all at once.
Strengths & weaknesses
Strengths
- Highest reasoning ceiling available
- Excellent tool use mid-reasoning
- Strong coding
Weaknesses
- The most expensive model in the catalogue
- Slow for interactive use
Pricing
| Rate | Price | Unit |
|---|---|---|
| Input | — quoted on request | — |
| Output | — quoted on request | — |
This model is not provisioned yet; the rate is quoted on request.
Context window
300Ktokens
covers prompt and response together, so a long input leaves less room for the answer.
Supported features
| Feature | Support |
|---|---|
| Tool calling | Supported |
| JSON mode | Supported |
| Streaming | Supported |
| Vision input | Not supported |
| Audio input | Not supported |
| Long context | Not supported |
| Open weights | Not supported |
| Reasoning | Supported |
| Fine-tunable | Not supported |
| Batch | Not supported |
| Caching | Not supported |
Example use cases
Hardest problems only
Scientific analysis
Verification of critical outputs
Comparison
Compare against anything →| Attribute | o4 | Claude Opus 4.5 | DeepSeek R1 |
|---|---|---|---|
| Context window | 300K | 200K | 164K |
| Input price | R0.00 | R0.00 | R0.00 |
| Output price | — not charged | — not charged | — not charged |
| Tool calling | Supported | Supported | Supported |
| JSON mode | Supported | Supported | Not supported |
| Vision input | Not supported | Supported | Not supported |
| Open weights | Not supported | Not supported | Supported |
| Prompt caching | Not supported | Supported | Not supported |
Best value in each row is highlighted. Blank cells are facts the supplier does not publish, not zeroes.
Provider information
View OpenAI →OpenAI · San Francisco, United States · est. 2015
OpenAI set the template the whole industry follows, from GPT's scaling curves to the API-first business model every provider now imitates. Its GPT-5 generation and o-series reasoning models remain the benchmark others measure against, and its developer platform is the most imitated interface in AI.
Recent releases
Documentation
Frequently asked questions
Input is billed at R0.00 per 1M tokens, and there is no separate output charge. Every figure on this page is an illustrative placeholder.
No. o4 is served through its provider's API only; the weights are not published, so self-hosting is not an option.
Prompt and response share one window: o4 accepts up to 300K tokens in total, so a long input leaves less room for the answer. The median across reasoning models is 200K. Recall is strongest well inside the limit. Test at your real prompt depth before relying on the ceiling.
Claude Opus 4.5 is the closest neighbour in the catalogue, with 200K of context at R0.00 per 1M tokens in. Put both in a comparison before wiring a fallback chain, so the trade-off is a decision rather than a default.