Migrating from gpt-5-nano to gpt-5.4-nano

What changes when you move from gpt-5-nano to gpt-5.4-nano: input price +300%, output price +212.5%.

From: gpt-5-nano
Status: Active
Shutdown: none announced
Time left: —

To: gpt-5.4-nano
Status: Deprecated
Shutdown: Apr 1, 2027
Time left: 168 days

gpt-5.4-nano is itself deprecated, shutting down Apr 1, 2027. Its replacement is gpt-6-luna.

Prices

USD per 1M tokens, standard tier

Type	gpt-5-nano	gpt-5.4-nano	Change
Input	$0.05first seen October 1, 2026 · source	$0.20first seen October 1, 2026 · source	+300%
Cached input	$0.005first seen October 1, 2026 · source	$0.02first seen October 1, 2026 · source	+300%
Output	$0.40first seen October 1, 2026 · source	$1.25first seen October 1, 2026 · source	+212.5%

At 1M input and 250K output tokens a month: $0.15 on gpt-5-nano, $0.5125 on gpt-5.4-nano (+241.7%).

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

Limits

Tokens

Limit	gpt-5-nano	gpt-5.4-nano	Change
Context window	400,000 tokensin place when tracking began (October 6, 2026) · source	400,000 tokensin place when tracking began (October 7, 2026) · source	0%
Max output tokens	128,000 tokensin place when tracking began (October 6, 2026) · source	128,000 tokensin place when tracking began (October 7, 2026) · source	0%

Benchmarks

Same benchmark, source and setting on both sides

Benchmark	gpt-5-nano	gpt-5.4-nano	Change
Arena TextArena, high effort. Ranges don't overlap.	1338 ±7	1401 ±4	+63
Arena Text: CodingArena, high effort. Ranges don't overlap.	1385 ±15	1460 ±6	+75
Arena VisionArena, high effort. Ranges don't overlap.	1146 ±11	1199 ±6	+53

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-nano: in place when tracking began (October 6, 2026) · source · gpt-5.4-nano: in place when tracking began (October 6, 2026) · source
Function calling: supported on both.gpt-5-nano: in place when tracking began (October 6, 2026) · source · gpt-5.4-nano: in place when tracking began (October 6, 2026) · source
Prompt caching: supported on both.gpt-5-nano: in place when tracking began (October 6, 2026) · source · gpt-5.4-nano: in place when tracking began (October 6, 2026) · source
Batch API: supported on both.gpt-5-nano: in place when tracking began (October 6, 2026) · source · gpt-5.4-nano: in place when tracking began (October 7, 2026) · source
Image input: supported on both.gpt-5-nano: in place when tracking began (October 6, 2026) · source · gpt-5.4-nano: in place when tracking began (October 7, 2026) · source
Reasoning / thinking: supported on both.gpt-5-nano: in place when tracking began (October 6, 2026) · source · gpt-5.4-nano: in place when tracking began (October 7, 2026) · source
Fine-tuning: not supported on both.gpt-5-nano: in place when tracking began (October 6, 2026) · source · gpt-5.4-nano: in place when tracking began (October 7, 2026) · source

Not published for gpt-5-nano: PDF input, Built-in web search, temperature parameter, top_p parameter, logprobs.

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

Aliases

All aliases

No aliases point at either model.

On cloud platforms

All platform dates

Model	Platform	Platform ID	Shutdown	Time left
gpt-5.4-nano	Microsoft Foundryversion 2026-03-17	gpt-5.4-nano	Sep 21, 2027Active	341 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

Change the model string from gpt-5-nano to gpt-5.4-nano.
Re-check costs: gpt-5.4-nano costs more than gpt-5-nano (input +300%, output +212.5%).

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