GPT-4.1

GPT-4.1 arrived on 14 April 2025 as an API-only family built around one number: a context window of 1,047,576 tokens, roughly eight times what GPT-4o offered. OpenAI positioned it for instruction following and tool calling rather than for the hardest reasoning problems — there is no reasoning step, so answers come back at conversational latency. The pricing was the quiet news. At $2 per million input tokens and $8 per million output, GPT-4.1 undercut GPT-4o ($2.5 and $10) while carrying eight times the context. Cached input costs $0.5 per million. Output is capped at 32,768 tokens and the knowledge cutoff is 1 June 2024. The model reads text and images and writes text, and it remains one of the few OpenAI models still open to fine-tuning. OpenAI now recommends starting with GPT-5 for complex work, and in April 2026 it scheduled a long list of legacy models for shutdown. GPT-4.1 was not on that list: as of August 2026 the base model and its mini sibling are still served, while the cheapest member of the family, GPT-4.1 nano, switches off on 23 October 2026.

▸Videos

No videos yet.

▸Reviews