OpenAI has officially made GPT‑4.1 available directly within ChatGPT – and it’s packed with significant upgrades, especially for developers, technical writers, and power users.
What’s New in GPT‑4.1?
Best-in-Class for Coding
- 54.6% on SWE-bench Verified (⬆️ 21.4% vs GPT‑4o)
- More reliable in frontend tasks, diff formatting, tool usage, and fewer unnecessary edits
Massive Context Window
- Handles up to 1 million tokens, ideal for long documents, codebases, legal texts, and customer support logs
Faster and Smarter
- GPT‑4.1 mini: ~2× faster with 83% lower cost
- GPT‑4.1 nano: blazing fast, ultra-cheap, ideal for classification and autocomplete
Superior Instruction Following
- 38.3% on MultiChallenge benchmark (10.5% vs GPT‑4o)
- Improved accuracy in format compliance, ranking, and confidence calibration
Multimodal Context Understanding
- New state-of-the-art result on Video-MME benchmark
GPT‑4.1 models are better suited than ever for powering AI agents – autonomous systems that complete tasks like bug fixing, email sorting, data analysis, and more.
Model Lineup
- GPT‑4.1 – Full model, long context, high performance
- GPT‑4.1 mini – Faster, smarter replacement for GPT‑4o mini
- GPT‑4.1 nano – Super low-latency, low-cost model for rapid tasks
GPT‑4.1 is now available in ChatGPT for Plus, Pro, and Team users via the ‘More Models’ dropdown. GPT‑4.5 Preview will be retired on 14th July 2025. Enterprise and Edu access is coming in the next few weeks.
