OpenAI releases GPT-4.1 family with 1M token context and coding improvements

OpenAI has released GPT-4.1, a new family of models designed for coding and instruction following, available via API since April 14 and in ChatGPT since May 14. The suite includes three sizes—GPT-4.1, Mini, and Nano—all supporting up to 1 million tokens of context.

• GPT-4.1 outperforms GPT-4o on SWE-bench (54.6% vs 33.2%) and instruction following benchmarks.
• GPT-4.1 Mini offers similar capabilities with lower latency and cost, becoming the default for free-tier users.
• GPT-4.1 Nano is the smallest and cheapest model at ~10 cents per million tokens, optimized for speed.
• All models support 1 million token context, an 8x increase over previous versions.

The release aims to provide developers with more predictable instruction following and efficient long-context processing while reducing costs compared to prior generations.