One agent session
1M input tokens and 100K output tokens, no caching.
- Input1M x $2 = $2.00
- Output0.1M x $8 = $0.80
$2.80per session
OpenAI - Closed weights
o3 pricing and specs
o3 is a closed-weights model from OpenAI that costs $2 per 1M input tokens and $8 per 1M output tokens on the first-party API, with a 200K-token context window.
Last verified from developers.openai.com. All sources
- Input / 1M
- $2
- Output / 1M
- $8
- Context
- 200K tokens
- Weights
- Closed
Pricing per 1M tokens
| Item | Price / 1M tokens |
|---|---|
| Input | $2 |
| Output | $8 |
| Cache read | $0.50 |
| Batch inputBatch requests | $1 |
| Batch outputBatch requests | $4 |
| Fast mode inputFast mode requests | $3.50 |
| Fast mode outputFast mode requests | $14 |
Note: Flex is $1 input / $0.25 cached / $4 output. Default snapshot o3-2025-04-16.
Specs
- Context window
- 200,000 tokens
- Max output
- 100,000 tokens
- Release date
- Provider
- OpenAI
- API model id
o3- Vercel AI Gateway id
openai/o3- Weights
- Closed weights
What o3 costs in practice
Worked from the standard list rates above, with the math shown so you can swap in your own token counts.
A chat app
10,000 requests a day at 2K input and 500 output tokens each.
- Input20M x $2 = $40.00
- Output5M x $8 = $40.00
$80.00per day
$2,400.00 per 30-day month
The same session at a 90% cache hit
90% of the input is read from the prompt cache. The one-time cache write premium is not included.
- Cached input0.9M x $0.50 = $0.45
- Uncached input0.1M x $2 = $0.20
- Output0.1M x $8 = $0.80
$1.45per session
Compare with
| Model | Input / 1M | Output / 1M |
|---|---|---|
| o3 (this page) | $2 | $8 |
| GPT-4.1OpenAI | $2 | $8 |
| Jamba Large 1.7AI21 Labs | $2 | $8 |
| Kimi K2.7 Code HighSpeedMoonshot AI | $1.90 | $8 |
| Qwen3.5-Omni-PlusAlibaba Cloud (Qwen) | $1.40 | $8.30 |
| GPT-4oOpenAI | $2.50 | $10 |
| GPT-5OpenAI | $1.25 | $10 |
| GPT-5.1OpenAI | $1.25 | $10 |
| GPT-6 SolOpenAI | $2 | $10 |
| GPT-6.1 SolOpenAI | $2 | $10 |
Related reading
- Codex CLI 0.163: Instant Steering Is Now the DefaultCodex CLI 0.163.0 makes mid-response steering the default and adds searchable /config and a leader key. What changed and what to check before upgrading.
- Tool Calling Under the Hood: Claude Messages vs OpenAI ResponsesHow tool calling works on the wire: JSON schema tools, parallel calls, tool_choice, strict mode and streaming deltas in Claude Messages vs OpenAI Responses.
- Codex 'gpt-5.5 Model Is Not Supported' Error: The FixCodex says gpt-5.5 is not supported with a ChatGPT account because OpenAI retired it from ChatGPT sign-in on October 14. The fix and what still works.
- Codex CLI Background Daemon: How to Turn It OffCodex CLI starts a persistent, self-updating background daemon since 0.157.0. What it does, why it hit Hacker News and the reported workaround to turn it off.
- Codex CLI 0.162: Managed Worktrees and Pinned TasksCodex CLI 0.162 lets the agent create and list its own Git worktrees, adds pinned tasks and custom provider capability overrides. What to change this week.
- ChatGPT Free vs Go vs Plus After GPT-6: Which Model You GetFree and Go ($8) both get GPT-6 Luna in Chat. Plus ($20) is the cheapest plan with GPT-6 Sol and full Codex. What each ChatGPT plan buys now.
FAQ
How much does o3 cost per million tokens?
o3 costs $2 per 1M input tokens and $8 per 1M output tokens on OpenAI's first-party API, per developers.openai.com on October 17, 2026. Per 1M tokens, cache reads cost $0.50, and the batch rate is $1 input and $4 output.
What is o3's context window?
o3 has a 200,000-token context window and returns up to 100,000 output tokens per response.
Are o3's weights open?
No. o3 is a closed-weights model. You call it through OpenAI's API as o3 or through Vercel AI Gateway as openai/o3.
When was o3 released?
OpenAI released o3 on April 16, 2025.
Sources
Last verified from developers.openai.com.
Newsletter
Get model price changes as they ship
One email per week. New model releases, price cuts and what they mean for your bill.