gpt-4o-mini
by openai
Pricing
Input
$0.06 / 1M tokens
Output
$0.24 / 1M tokens
Overview
GPT-4o Mini is a compact version of GPT-4o, launched in July 2024. It offers 128K context, image input, and GPT-4-level reasoning at a fraction of the cost.
- Release Date: July 18, 2024
- Multimodal Support: Text + Images (Audio/Video coming soon)
- Context Length: 128,000 tokens
- Output Limit: Up to 16,000 tokens
Features
- Lower latency and lightweight
- Multilingual capabilities
- Strong math, coding, and logic performance
- Vision support via image input
- Up to 60% cheaper than GPT-3.5 Turbo
Benchmarks
| Task | GPT-4o Mini | Gemini Flash | Claude Haiku |
|---|---|---|---|
| MMLU (reasoning) | 82.0% | 77.9% | 73.8% |
| MGSM (math) | 87.0% | 79.7% | 67.8% |
| HumanEval (coding) | 87.2% | 74.1% | 64.9% |
| MMMU (multimodal) | 59.4% | 56.8% | 54.9% |
Pricing
| Type | Price per 1M tokens |
|---|---|
| Input | $0.15 |
| Output | $0.60 |
60% cheaper than GPT-3.5 Turbo with better reasoning and generation.
Use Cases
- Real-time chatbots and assistants
- Study help and educational tools
- Document processing and summarization
- Lightweight embedded LLMs
- OCR, diagram, and image-based Q&A apps
Safety and Stability
- Built on GPT-4o safety framework
- Instruction-following and jailbreak resistance
- No web browsing (offline knowledge)
- Knowledge cutoff: October 2023
Limitations
- Inconsistencies in numeric/math extraction
- Audio/video not yet supported
- No live internet connection (unless paired with search plugin)
License
Commercial usage permitted under OpenAI's terms. GPT-4o Mini can be embedded in SaaS products, enterprise tools, and developer platforms.