AI model timeline
Releases, price changes, deprecations and retirements across every provider, as they happen.
Coming up
DeepSeek starts billing its Search and Fetch APIs per request
From 00:00 Beijing time on 1 November 2026, standalone Search and Fetch calls cost $0.75 per 1,000 requests. Calls made through DeepSeek Harness, Claude Code or Codex are currently free.
SourceGoogle publishes higher Gemini API prices for December 2026 and January 2027
Gemini 4 Argon prices double from 1 December 2026, and most other Gemini models double from 1 January 2027, on the standard paid tier. The changes are already shown as scheduled on each model page here.
Source
October 2026
Claude API version 2026-10-21 removes temperature, top_p and top_k
Requests that set these sampling parameters return an error on the new API version. Anthropic recommends steering behaviour through prompting instead. Earlier API versions already rejected non-default values on Claude 4.7 and later models.
Source
September 2026
Claude Sonnet 5's $2 / $10 launch price becomes its standard price
Introduced as launch pricing through 31 August 2026, the $2 input and $10 output per million tokens is now permanent. The increase to $3 / $15 scheduled for 1 September 2026 will not happen.
Source
July 2026
OpenAI renames Priority processing to the Fast tier
OpenAI's higher-speed processing option is now called the Fast tier, alongside the Standard, Batch, Flex and Ultrafast tiers.
Source