All times Eastern (ET)Provider pages checked daily at
ProSign inRSS

Migrating from gpt-5-mini to gpt-5.4-mini

What changes when you move from gpt-5-mini to gpt-5.4-mini: input price +200%, output price +125%.

Status
Active
Shutdown
none announced
Time left
—
Status
Active
Shutdown
none announced
Time left
—

Prices

USD per 1M tokens, standard tier
Typegpt-5-minigpt-5.4-miniChange
Input$0.25first seen October 1, 2026 · source$0.75first seen October 1, 2026 · source+200%
Cached input$0.025first seen October 1, 2026 · source$0.075first seen October 1, 2026 · source+200%
Output$2.00first seen October 1, 2026 · source$4.50first seen October 1, 2026 · source+125%

At 1M input and 250K output tokens a month: $0.75 on gpt-5-mini, $1.875 on gpt-5.4-mini (+150%).

Other tiers, prompt-size bands and modalities are on each model's page.

Limits

Tokens
Limitgpt-5-minigpt-5.4-miniChange
Context window400,000 tokensin place when tracking began (October 6, 2026) · source400,000 tokensin place when tracking began (October 6, 2026) · source0%
Max output tokens128,000 tokensin place when tracking began (October 6, 2026) · source128,000 tokensin place when tracking began (October 6, 2026) · source0%

Benchmarks

Same benchmark, source and setting on both sides
Benchmarkgpt-5-minigpt-5.4-miniChange
Arena TextArena, high effort. Ranges don't overlap.1390 ±51447 ±4+57
Arena Text: CodingArena, high effort. Ranges don't overlap.1432 ±81495 ±6+63
Arena VisionArena, high effort. Ranges don't overlap.1183 ±81251 ±6+68

A score from one source is never compared with another source's, and company-reported scores are compared only within the same company; settings named where they differ. Data: Arena (arena.ai), CC BY 4.0.

Features and parameters

Compare models

Same on both

  • Structured outputs: supported on both.gpt-5-mini: in place when tracking began (October 6, 2026) · source · gpt-5.4-mini: in place when tracking began (October 6, 2026) · source
  • Function calling: supported on both.gpt-5-mini: in place when tracking began (October 6, 2026) · source · gpt-5.4-mini: in place when tracking began (October 6, 2026) · source
  • Batch API: supported on both.gpt-5-mini: in place when tracking began (October 6, 2026) · source · gpt-5.4-mini: in place when tracking began (October 6, 2026) · source
  • Image input: supported on both.gpt-5-mini: in place when tracking began (October 6, 2026) · source · gpt-5.4-mini: in place when tracking began (October 6, 2026) · source
  • Reasoning / thinking: supported on both.gpt-5-mini: in place when tracking began (October 6, 2026) · source · gpt-5.4-mini: in place when tracking began (October 6, 2026) · source
  • Fine-tuning: not supported on both.gpt-5-mini: in place when tracking began (October 6, 2026) · source · gpt-5.4-mini: in place when tracking began (October 6, 2026) · source
  • Built-in web search: supported on both.gpt-5-mini: in place when tracking began (October 6, 2026) · source · gpt-5.4-mini: in place when tracking began (October 6, 2026) · source

Not published for gpt-5-mini: Prompt caching, PDF input, temperature parameter, top_p parameter, logprobs.

Not published for gpt-5.4-mini: PDF input, temperature parameter, top_p parameter, logprobs.

Aliases

All aliases

No aliases point at either model.

On cloud platforms

All platform dates
ModelPlatformPlatform IDShutdownTime left
gpt-5.4-miniMicrosoft Foundryversion 2026-03-17gpt-5.4-miniSep 21, 2027Active341 days

Read daily from each platform's own pages. The status above uses each company's own date. Time left counts to 00:00 UTC on the shutdown date, which is the evening before in ET.

Checklist

  1. Change the model string from gpt-5-mini to gpt-5.4-mini.
  2. Re-check costs: gpt-5.4-mini costs more than gpt-5-mini (input +200%, output +125%).

Compared from AIWire's published data when this page loaded. Each fact links to the page it came from. Model pages: gpt-5-mini, gpt-5.4-mini.