- Start
- May 14, 202595% CONFIDENCEfrom help.openai.com
GPT-4.1 Comes to ChatGPT
- GPT-4.1 reached ChatGPT for paid users on May 14, 2025, alongside GPT-4.1 mini replacing GPT-4o mini[1]
- It had launched in the API on April 14, 2025 with GPT-4.1 mini and nano, up to 1 million tokens of context and a June 2024 knowledge cutoff; GPT-4.1 itself was API-only until then, with improvements going gradually into the latest GPT-4o in ChatGPT[2]
- GPT-4.5 Preview was deprecated in the API, to be turned off July 14, 2025[2]
Notable features
- A GPT model family that outperforms GPT-4o and GPT-4o mini across the board, with major gains in coding, instruction following and long-context comprehension over up to 1 million tokens[3]
- API results: SWE-bench Verified 54.6% (GPT-4o 33.2%), IFEval 87.4% vs 81.0%, extraneous code edits down from 9% to 2%[2]
- On Scale's MultiChallenge instruction-following benchmark it scores 38.3%, 10.5 points above GPT-4o, and on Video-MME (long, no subtitles) 72.0% against 65.3%[3]
- API pricing: $2.00 input and $8.00 output per million tokens, 26% cheaper than GPT-4o for median queries[2]
- Long-context requests cost nothing beyond standard per-token prices and the prompt caching discount rose to 75%; in initial testing time to first token was about 15 seconds at 128,000 tokens of context and a minute at a million[3]
References 390% CONFIDENCE
The first entry is always the pin's source. Overall confidence is a weighted average of how firmly each reference supports the start and end times used above; a reference counts half as much for every 180 days older than the newest.
- [1]95%OpenAI Model Release Notes: Releasing GPT-4.1 in ChatGPT for all paid users (May 14, 2025)help.openai.com· Posted Sep 19, 2026· Starts May 14, 2025· 47% of score
OpenAI's[2] own model release notes date this entry May 14, 2025
- [2]85%Introducing GPT-4.1 in the APIopenai.com· Added Sep 19, 2026· Starts May 14, 2025· 47% of score
OpenAI's[1] announcement post, linked from the release note entry.
- [3]88%Introducing GPT-4.1 in the API | OpenAI (Wayback Machine copy)web.archive.org· Published Apr 14, 2025· 6% of score
Archived copy of OpenAI's[1][2] GPT-4.1 announcement, which gives MultiChallenge 38.3% ("a 10.5% abs increase over GPT-4o"), Video-MME 72.0% "up from 65.3% for GPT-4o", and a prompt caching discount of "75% (up from 50% previously)".
Suggest a correction
Something missing or wrong? Say it in your own words: a link that backs this pin up, a different start or end date and why, or a fact it lacks or gets wrong. The AI checks it against this pin's sources, searches for better ones, and adds any page that backs you up. The pin's own sources still count most.
