GPT-4o mini
Language Available Full comparison ↗ openai/gpt-4o-mini · by OpenAI
· decoder-only-transformer
Pricing — 1 offering(s)
About this offering: OpenAI GPT-4o mini. Small, fast, cost-optimised sibling of GPT-4o. Widely used as a cheap default for high-volume classification, extraction, and simple chat workloads. 128K context window.
Input tokens
- $0.15 / 1M tokens (input) Current 2024-07-18 → present
Output tokens
- $0.60 / 1M tokens (output) Current 2024-07-18 → present
Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.
See how GPT-4o mini fits into a cost-aware routing setup
See how →Capability profile
How GPT-4o mini rates across core capability dimensions, with the task-level evidence behind each rating.
Solid for structured tasks; significantly below GPT-4o on hard multi-step problems.
Good at boilerplate and simple scripts; struggles with complex refactoring.
Reliable function calling; widely used in tool-calling production pipelines.
Consistent format compliance and system prompt adherence.
128K context.
Strong multilingual coverage inherited from GPT-4o training.
Low-latency; designed for high-volume applications.
Operator guidance
The default choice for cost-sensitive OpenAI workloads. Route here first for anything that doesn't demonstrably require GPT-4o reasoning depth. Step up to GPT-4o or GPT-5.5 when task quality suffers; step across to Llama 4 Maverick for the same price with open weights.
Use cases
- High-volume classification, extraction, and simple chat workloads
- Tool-calling pipelines where cost must be minimised
- Lightweight RAG reranking and summarisation
Limitations
- Noticeably weaker than GPT-4o on multi-hop reasoning and complex coding
- Capability ratings are qualitative, not a single cited benchmark run