GPT-4 Turbo

GPT-4 Turbo reached general availability on 9 April 2024, seventeen months after the preview build that OpenAI had shown in November 2023 (the earlier gpt-4-turbo-preview line dates to 25 January 2024). It was designed as a cheaper and better GPT-4: a 128,000-token context window against the original 8,192, image input, and a knowledge cutoff of 1 December 2023. The output limit is the number that dates it. GPT-4 Turbo returns at most 4,096 tokens per response — roughly a thirtieth of what the GPT-5 generation allows — so long documents had to be assembled across several calls. Pricing was $10 per million input tokens and $30 per million output, with no cached-input rate published; that is eight times what GPT-4o mini charges for input three months later, and it explains how quickly the omni family displaced it. OpenAI has scheduled the shutdown for 23 October 2026, in the same batch as GPT-3.5 Turbo, GPT-4, GPT-4.1 nano, o1 and o1-pro, with GPT-5.6 Sol named as the substitute. The profile stays in this catalogue after that date; the entry for the original GPT-4 is kept separately.

▸Videos

No videos yet.

▸Reviews