AI Radar

GPT-5.4 mini

Available · Language, Multimodal, Reasoning

GPT-5.4 mini is a small OpenAI model for text and image input, released on 17 March 2026 with GPT-5.4 nano for high-volume workloads, in the API, Codex and ChatGPT. OpenAI says it brings many strengths of GPT-5.4 to a faster model. GPT-5.4 nano was offered only in the API and was deprecated on 1 October 2026. [1] [2]Model details [3]2026-10-01: GPT-5.3-Codex, GPT-5.1, GPT-5.4-Nano Primary source

Timeline of GPT-5.4 mini →

Claims and evidence

  • Released Release in the API and Codex, and in ChatGPT per the announcement; the ChatGPT release notes date the ChatGPT rollout one day later. Primary source[1]Availability & pricing [4]Mar 17 (2026)
  • Rolled out in ChatGPT Primary source[5]March 18, 2026: GPT-5.4 mini in ChatGPT [6]GPT-5.4 mini in ChatGPT (March 18, 2026)
  • Status AvailableGPT-5.4 mini is listed in the API model catalog and not on the deprecations page; its nano variant was deprecated on 2026-10-01. Primary source[2] [3]
  • Successor of GPT-5The post presents GPT-5.4 mini as an improvement over GPT-5 mini and GPT-5.4 nano as an upgrade over GPT-5 nano; both are inline variants of the GPT-5 record. Primary source[1]opening paragraphs
  • Change · Efficiency OpenAI says it runs more than twice as fast as GPT-5 mini while improving coding, reasoning, multimodal understanding and tool use.Compared with GPT-5 Primary source[1]
  • Change · Tool use Supports tool search and built-in computer use in the API; GPT-5.4 nano supports neither.Compared with GPT-5 Primary source[4]Mar 17 (2026)
  • Change · Reasoning Reasoning effort extends to xhigh; for GPT-5 mini the top setting was high.Compared with GPT-5 Primary source[1] [2]
  • Input text, image Primary source[2]Model details [1]Availability & pricing
  • Output text Primary source[2]Model details
  • Feature Reasoning modeReasoning effort none (default), low, medium, high and xhigh. Primary source[2]
  • Feature Computer use Primary source[1]Availability & pricing [4]Mar 17 (2026)
  • Feature Function calling Primary source[1]Availability & pricing
  • Context window 400,000 tokensThe announcement says 400k; the docs list a maximum of 272,000 input tokens. Primary source[2]Model details
  • Parameters Not disclosedNot disclosed by OpenAI. No claim madeNo source
  • Access API, consumer appIn ChatGPT it is not in the model picker: Free and Go users reach it through the Thinking feature, other users as a rate-limit fallback for GPT-5.4 Thinking. Primary source[1]Availability & pricing

Lineage

Predecessors

Successors

Based on

Not derived from another model.

Variants and derived

None recorded.

Siblings

None recorded.

All ancestors

All descendants

Variants

GPT-5.4 nano

  • Released Primary source[1]Availability & pricing [4]Mar 17 (2026)
  • Deprecated Primary source[3]2026-10-01: GPT-5.3-Codex, GPT-5.1, GPT-5.4-Nano
  • Variant Inline variant in this record.Smallest and cheapest GPT-5.4-class model, released the same day in the API only. Supports compaction but not tool search or computer use (changelog). Same 400,000-token context window as GPT-5.4 mini. Primary source[1] [7]

Status: Deprecated Evidence not assessed [3]2026-10-01: GPT-5.3-Codex, GPT-5.1, GPT-5.4-Nano

Also known as: gpt-5.4-nano, gpt-5.4-nano-2026-03-17

Differs in:

  • Access: API Evidence not assessed [1]Availability & pricing

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from OpenAI on AI Radar →