Cheaper Inference
Model record / OpenAI
o3-fast is available for text generation from OpenAI. Its reported context window is 200,000 tokens, with up to 100,000 output tokens. Reported inputs: image, pdf, text. The served catalog lists image input, streaming.
Ready when you are
Marketplace pricing, capped at list, optimized further.
POST /v1/chat/completions
curl https://api.cheaperinference.com/v1/chat/completions \ -H "Authorization: Bearer ci_live_YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "o3-fast", "messages": [ { "role": "user", "content": "Reply with the single word OK." } ] }'POST /v1/responses
curl https://api.cheaperinference.com/v1/responses \ -H "Authorization: Bearer ci_live_YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "input": [ { "role": "user", "content": "Reply with the single word OK." } ], "model": "o3-fast", "store": false }'POST /v1/messages
curl https://api.cheaperinference.com/v1/messages \ -H "Authorization: Bearer ci_live_YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "o3-fast", "messages": [ { "role": "user", "content": "Reply with the single word OK." } ], "max_tokens": 1024 }'Capability source · checked
$5.95 estimated at displayed rates
1,000,000 ÷ 1,000,000 × $2.975 + 250,000 ÷ 1,000,000 × $11.90
The same billed quantities at the listed reference rates: $7.00.
Uses uncached input at the displayed standard rates, with no extra modalities or tools. Billed output includes any charged reasoning tokens. Actual token counts, price tiers, and marketplace prices can change the total.
Price quote effective .
Python: pip install openai · Node.js: npm install openai. Set CHEAPER_INFERENCE_API_KEY on your server.
import osfrom openai import OpenAIclient = OpenAI( api_key=os.environ["CHEAPER_INFERENCE_API_KEY"], base_url="https://api.cheaperinference.com/v1",)request = { "input": [ { "role": "user", "content": "Reply with the single word OK.", }, ], "model": "o3-fast", "store": False,}response = client.responses.create(**request)print(response.output_text)import OpenAI from "openai";const client = new OpenAI({ apiKey: process.env.CHEAPER_INFERENCE_API_KEY, baseURL: "https://api.cheaperinference.com/v1",});const response = await client.responses.create({ "input": [ { "role": "user", "content": "Reply with the single word OK." } ], "model": "o3-fast", "store": false});console.log(response.output_text);For other supported endpoints and their exact parameters, use the request samples and parameter reference on this page. Read the API integration guide.
| Model | Current pricing | Context limit | Image input listed |
|---|---|---|---|
| o3-fast (this model)OpenAI | $2.975 input / $11.90 output per 1M tokens | 200,000 tokens | Yes |
| gpt-5-fastOpenAI | $2.125 input / $17.00 output per 1M tokens | 400,000 tokens | Yes |
| gpt-5-mini-fastOpenAI | $0.3825 input / $3.06 output per 1M tokens | 400,000 tokens | Yes |
| gpt-5.1-thinking-fastOpenAI |
POST /v1/chat/completions
| Field | Values |
|---|---|
| reasoning.exclude | |
| max_tokens | |
| reasoning | effort: low, medium, high |
| response_format | |
| stop | |
| tool_choice | |
| tools |
POST /v1/messages
| Field | Values |
|---|---|
| max_tokens | |
| thinking | type: enabled with budget_tokens; output_config.effort: low, medium, high |
| output_config.format | |
| stop_sequences | |
| tool_choice | |
| tools |
POST /v1/responses
| Field | Values |
|---|---|
| max_output_tokens | |
| reasoning | effort: low, medium, high |
| text.format | |
| tool_choice | |
| tools |
See how much this model saves and when its recorded price has been lowest.
This model averaged 20.2% below comparable direct provider prices across the selected range. Its strongest observed window was 2:00 PM–7:00 PM in UTC.
Average saved
20.2%
Cheapest time
2:00 PM–7:00 PM
Hourly coverage
23/24
Signal quality
Growing
Hover a bar for savings per token and observation counts.
Source buckets are UTC; labels are converted to UTC.
Last 24 hours
#242
of 332 ranked models
Last 7 days
#295
of 389 ranked models
Last 30 days
#263
of 400 ranked models
| $2.125 input / $17.00 output per 1M tokens |
| 400,000 tokens |
| Yes |