MultimodalAPI only
GPT-4o
OpenAI
Updated May 21, 2026 · Benchmarks via Artificial Analysis, specs & pricing via OpenRouter · Methodology · Report an error
Superseded by GPT-4.1
Specifications
- Type
- Multimodal
- Access
- API only
- Released
- May 13, 2024
- License
- Proprietary
- Context window
- 128,000 tokens
- Knowledge cutoff
- October 31, 2023
- Output speed
- 132 tok/s
- Latency (TTFT)
- 0.5s
- Input
- Text, Image
- Output
- Text
- API pricing
- $2.50 in · $10.00 out / 1M tokens
Capabilities
Function callingStructured output
Benchmarks
Reasoning
| GPQA Diamond | 70.1 |
Coding
| HumanEval | 90.2 |
| Aider Polyglot | 30.7 |
| SWE-bench Verified | 33.2 |
| Aider Polyglot Edit | 18.2 |
| LiveCodeBench | 42.5 |
Agents
| TAU-bench Airline | 42.8 |
| TAU-bench Retail | 60.3 |
| τ²-bench Airline | 45.5 |
| τ²-bench Retail | 63.4 |
Our take
OpenAI’s first natively multimodal flagship — text, vision, and audio in a single model with sharply lower latency and price than GPT-4 Turbo. It reset the default price/performance bar that competitors chased through 2024.