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

GPT-5.4 mini

API ID gpt-5.4-mini

gpt-5.4-mini is an OpenAI model, released March 2026. It has a 400,000-token context window and up to 128,000 output tokens. It takes image input and supports function calling, structured outputs and reasoning. Standard price: from $0.75 input and from $4.50 output per 1M tokens.

Status
Active
Released
March 2026
Shutdown
none announced
Replacement
—

No deprecation or shutdown date has been announced as of October 15, 2026.

Benchmarks

Scores as each source published them
or

Comparing with GPT-5 Mini, the previous model in this line. A score is compared only with the same benchmark from the same source, at the same setting where both have it. Full comparison: prices, limits, features

Arena

ArenaGPT-5.4 miniGPT-5 MiniChangeRankDateSource
Text
By category
  • Math 1439 rank 93 vs 1407 (+32)
  • Creative writing 1402 rank 107 vs 1323 (+79)
  • Instruction following 1434 rank 90 vs 1373 (+61)
  • Hard prompts 1469 rank 84 vs 1404 (+65)
  • Long prompts 1447 rank 96 vs 1371 (+76)
  • Multi-turn 1466 rank 70 vs 1373 (+93)
  • Expert 1481 rank 73 vs 1402 (+79)
1447 ±41390 ±5+5784 of 415
63,461 votes		Oct 13, 2026source
Text: Coding1495 ±61432 ±8+6388 of 410
17,247 votes		Oct 13, 2026source
WebDev1398 ±7not rated93 of 142
11,978 votes		Oct 13, 2026source
Vision1251 ±61183 ±8+6853 of 165
25,733 votes		Oct 9, 2026source

Rated as gpt-5.4-mini-high (high effort); GPT-5 Mini as gpt-5-mini-high (high effort). Ratings come from people voting between two anonymous answers; Text and Vision are style-controlled, as on arena.ai. Data: Arena (arena.ai), CC BY 4.0.

Epoch AI

BenchmarkGPT-5.4 miniGPT-5 MiniChangeSettingRunSource
FrontierMath Tier 49.8%not runxhigh effortJun 12, 2026source
FrontierMath Tiers 1-3
Other settings
  • 24.6% vs not run low effort
  • 17.2% vs not run no reasoning
51.2%not runxhigh effortJun 12, 2026source
GPQA Diamond
Other settings
  • 83.6% vs not run high effort
  • 64.1% vs not run no reasoning
86.9%not runxhigh effortAug 7, 2026source
OTIS Mock AIME 2024-2025
Other settings
  • 87.2% vs not run high effort
  • 26.7% vs not run no reasoning
88.9%not runxhigh effortAug 7, 2026source
SimpleQA Verified29.4%not runhigh effortAug 27, 2026source

Run independently by Epoch AI. Data: Epoch AI, Capabilities & benchmarking (epoch.ai/benchmarks), CC BY.

On cloud platforms

All platform dates
PlatformPlatform IDStateShutdownTime leftReplacementSource
Microsoft Foundryversion 2026-03-17gpt-5.4-miniActiveSep 21, 2027341 days—source

Read daily from each platform's own pages. The shutdown date and countdown on this page use OpenAI's own date. Time left counts to 00:00 UTC on the shutdown date, which is the evening before in ET.

Alias

All aliases

Currently points to gpt-5.4-mini-2026-03-17 (in place when tracking began (October 7, 2026)).

Features

Compare models
FeatureValueSinceSource
Structured outputsThe model can be held to a JSON schema you supply, so every reply parses without retries or cleanup.Supportedin place when tracking began (October 6, 2026)source
Function callingAlso called tool use. You describe functions; the model replies with a function name and arguments for your code to run, then uses the result.Supportedin place when tracking began (October 6, 2026)source
Prompt cachingA prompt prefix you send again and again (system prompt, documents, tools) is cached, so later requests cost less and start faster.Supportedin place when tracking began (October 6, 2026)source
Batch APISend many requests as one job that runs in the background, usually finished within 24 hours, at a lower price than live requests.Supportedin place when tracking began (October 6, 2026)source
Image inputImages such as screenshots, photos and charts can go in the prompt for the model to read.Supportedin place when tracking began (October 6, 2026)source
Reasoning / thinkingThe model can work through a problem step by step before it answers: usually better answers to hard questions, at the cost of more tokens and time.Supportedin place when tracking began (October 6, 2026)source
Fine-tuningYou can train your own version of the model on your examples through the company's API, then call it like any other model.Not supportedin place when tracking began (October 6, 2026)source
Built-in web searchThe company's API can search the web for the model during a request and cite its sources; you don't need a search service of your own.Supportedin place when tracking began (October 6, 2026)source
Context windowThe most the model can work with in one request, in tokens: your prompt, conversation history and files. A token is about three-quarters of an English word.400,000 tokensin place when tracking began (October 6, 2026)source
Max output tokensThe longest reply the model can write in one response, in tokens. A token is about three-quarters of an English word.128,000 tokensin place when tracking began (October 6, 2026)source

Not published by OpenAI: PDF input, temperature parameter, top_p parameter, logprobs.

Price history

USD per 1M tokens, standard tier
$0.00$1.00$2.00$3.00$4.00$5.00Oct 2026In $0.75Out $4.50Oct 1, 2026 to Oct 15, 2026: Input $0.75, Output $4.50
$0.00$1.00$2.00$3.00$4.00$5.00Oct 2026In $0.75Out $4.50Oct 1, 2026 to Oct 15, 2026: Input $0.75, Output $4.50

We started recording this model's prices on Oct 1, 2026. The history grows with each daily check.

Show as a table
FromTypePrice
Oct 1, 2026Input$0.75
Oct 1, 2026Output$4.50

Prices

USD per 1M tokens, every type and band, by tier
TypeBandPrice $/1MIn effect
Standard
inputshort context$0.75
cached inputshort context$0.075
outputshort context$4.50
Batch
inputshort context$0.375
cached inputshort context$0.0375
outputshort context$2.25
Flex
inputshort context$0.375
cached inputshort context$0.0375
outputshort context$2.25
Fast
inputshort context$1.50
cached inputshort context$0.15
outputshort context$9.00

Changelog

Changes we detected on OpenAI's pages

No changes recorded yet.

Email alerts

Get an email if gpt-5.4-mini changes

We'll email you the same morning we spot a deprecation notice, a new or moved shutdown date, a price change, or the model leaving OpenAI's model list.

No password: one email to confirm, and you can unsubscribe any time. Privacy

Last checked · Sources: OpenAI deprecations, OpenAI pricing, OpenAI models