Scheduled for retirement on Nov 19, 2026. Still available at the prices below until then.
Context
200K
Max output
100K
Serving providers
4
Cheapest input
$2.00 /1M
On-demand listed rates, OpenRouter routed quotes and Vercel reseller quotes. Enterprise commitments (committed-use discounts, provisioned throughput, negotiated rates) can be materially lower.
Benchmarks
independent evaluations · source: Artificial AnalysisSpeed & latency
Output speed
121 t/s
tokens / second
Time to first token
6.79s
latency
Headline indices
Individual benchmarks · click a row for effort variants
| Benchmark | Best score | Effort | Cost / task | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CritPt | 1% | default | — | |||||||||||||
| GPQA Diamond | 83% | default | — | |||||||||||||
| ||||||||||||||||
| Humanity's Last Exam | 20% | default | — | |||||||||||||
| ||||||||||||||||
| IFBench | 71% | default | — | |||||||||||||
| LiveCodeBench | 81% | default | — | |||||||||||||
| Long-Context Reasoning | 75% | default | — | |||||||||||||
| MMLU-ProFrozen Aug 18, 2026 | 85% | default | — | |||||||||||||
| MMMU-Pro | 70% | default | — | |||||||||||||
| Omniscience | -15.6 | default | — | |||||||||||||
| Omniscience Accuracy | 0.4 | default | — | |||||||||||||
| Omniscience Non Hallucination | 0.1 | default | — | |||||||||||||
| SciCodeArtificial Analysis did not report this score in its latest run on Oct 14, 2026 | 41% | — | — | |||||||||||||
| TerminalBench Hard | 37% | default | — | |||||||||||||
| τ²-bench | 81% | default | — | |||||||||||||
Artificial Analysis last published scores for this benchmark on Aug 18, 2026. Scores here are frozen at that date, and models released since are not on it.
Pricing by serving provider
standard tier · per 1M tokens · click a provider to expand| Serving provider | Input /1M | Output /1M | Endpoints | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Azure | $2.00 | $8.00 | 1 | |||||||||||||||||
Standardazure/o3
| ||||||||||||||||||||
| OpenAICheapest | $2.00 | $8.00 | 1 | |||||||||||||||||
Standardo3
| ||||||||||||||||||||
| OpenRouter | $2.00 | $8.00 | 1 | |||||||||||||||||
Standardopenai/o3
| ||||||||||||||||||||
| VercelVercel reseller quote | $2.00 | $8.00 | 1 | |||||||||||||||||
Standardopenai/o3
| ||||||||||||||||||||
| OpenRouter (batch) | $1.00 | $4.00 | 1 | |||||||||||||||||
Batchopenai/o3:batch
| ||||||||||||||||||||
| Vercel (flex)Vercel reseller quote | $1.00 | $4.00 | 1 | |||||||||||||||||
Flexopenai/o3
| ||||||||||||||||||||
| Vercel (fast)Vercel reseller quote | $3.50 | $14.00 | 1 | |||||||||||||||||
Fastopenai/o3-fast
| ||||||||||||||||||||
| Vercel (priority)Vercel reseller quote | $3.50 | $14.00 | 1 | |||||||||||||||||
Priorityopenai/o3
| ||||||||||||||||||||
Price history
listed seller prices, input and output per 1M tokens Input Output
Input down 80% since first tracked
Cost calculator
estimate your monthly spend on this model$1,040
estimated / month
Model IDs
copy the exact identifier for your platformopenai/o3:batchopenai/o3azure/o3openai/o3o3openai/o3openai/o3-fastopenai/o3
Frequently asked questions
o3 pricing, context and availabilityIs o3 still available?
Yes. o3 is scheduled for retirement on 2026-11-19.
How much does o3 cost?
o3 costs $2.00 per 1M input tokens from OpenAI, and $8.00 per 1M output tokens from OpenAI.
What is the context window of o3?
o3 accepts up to 200K tokens of context and can return up to 100K output tokens.
Which providers serve o3?
o3 is available from 4 serving providers, each with its own pricing and model ID. OpenAI is currently the cheapest.
What can o3 do?
o3 supports image input (vision), tool use, reasoning, prompt caching, structured output and web search.
Other OpenAI models
compare pricing across the OpenAI lineupPopular comparisons
head-to-head pages featuring o3
o3 vs Claude 4 Opus Listed prices from LiteLLM, OpenRouter and Vercel · last checked Oct 14, 2026
Every weekday
AI moves fast. Here's your debrief.
News, analysis, tools, and more.
For people who build with AI