GPT-5.1 in the API, the next model in the GPT-5 series built to balance intelligence and speed across agentic and coding tasks.
Here’s what’s new in GPT 5.1:
- Adaptive reasoning that adjusts thinking time by task complexity—spending more time on complex tasks and responding faster on simple tasks.
- New reasoning_effort = 'none' mode, offers a fast, accurate non-reasoning path for latency sensitive use cases. Defaults to none when unspecified.
- Extended prompt caching with retention up to 24 hours to reduce latency for long-running conversations.
- Upgrades to coding: more communicative, highly steerable, better code quality, improved frontend UI generation.
- New tools: apply_patch (structured diffs) and shell (controlled local CLI).
- Pricing: Same pricing and rate limits as GPT-5
Codex Models: We also shipped gpt-5.1-codex and gpt-5.1-codex-mini. Here’s how to think about these models:
- gpt-5.1: for everyday coding tasks
- gpt-5.1-codex: for complex, long-running agentic coding
- gpt-5.1-codex-mini: for cost-efficient edits and changes
Next Steps:
We recommend you test GPT-5.1 for coding and consider switching from GPT-4.1 to GPT-5.1 on reasoning_effort='none' for fast general-purpose tasks.
Start building with GPT-5.1 and the latest examples in our API docs:
The OpenAI Team