Every line, version by version

GPT-5 versions compared

Every GPT-5 version in one table: input, output and cached prices, context window, benchmark scores, and what a month of your usage costs on each one.

The short version

  • GPT-5.6 cyber costs 18650% more than GPT-5 nano per output token
  • GPT-5 nano is the cheapest version in this line at $0.225 per million blended tokens
  • 3 of 5 versions in this line are now legacy models

Every GPT-5 version in one table

Prices are read live from the provider pricing pages, so this table cannot go stale.

VersionTierStatusReleasedInputOutputCachedContextBlended /1MSample /mo
GPT-5.6 cyberLatest
5.6
Price your own usage		FlagshipActive—$12.50$75.00$1.25400K$43.75$50.00
GPT-5.6 sol
5.6
Promotional pricing ($4 / $20 / $0.40) is available at least through 2026-11-21; the pre-promo rate was $5 / $30. Re-check after that date.
Price your own usage		FlagshipActive—$4.00$20.00$0.4001,050K$12.00$23.00Long context ≥272,000
GPT-5
5
Replaced by GPT-6 series; kept for long-tail search.
Price your own usage		FlagshipLegacy—$1.25$10.00$0.125400K$5.63$6.25
GPT-5 mini
5
Replaced by GPT-6 series; kept for long-tail search.
Price your own usage		FastLegacy—$0.250$2.00$0.025400K$1.13$1.25
GPT-5 nanoCheapest
5
Replaced by GPT-6 series; kept for long-tail search.
Price your own usage		LiteLegacy—$0.050$0.400$0.0050400K$0.225$0.250

Sample cost = 1M input + 500K output tokens per month at list price. A model with a long-context band is priced on that band — the rule is printed in the row. Use the calculator below for your own workload.

What changed at each step

Only the differences we can prove from the pricing and score records — a price that moved, a context window that grew, a benchmark both versions report.

  1. GPT-5.6 sol→GPT-5.6 cyber
    • Input price up 213% ($4.00 → $12.50)
    • Output price up 275% ($20.00 → $75.00)
    • Cached price up 213% ($0.400 → $1.25)
    • Context window down (1,050K → 400K)
  2. GPT-5→GPT-5.6 sol
    • Input price up 220% ($1.25 → $4.00)
    • Output price up 100% ($10.00 → $20.00)
    • Cached price up 220% ($0.125 → $0.400)
    • Context window up (400K → 1,050K)
  3. GPT-5 mini→GPT-5
    • Input price up 400% ($0.250 → $1.25)
    • Output price up 400% ($2.00 → $10.00)
    • Cached price up 400% ($0.025 → $0.125)
  4. GPT-5 nano→GPT-5 mini
    • Input price up 400% ($0.050 → $0.250)
    • Output price up 400% ($0.400 → $2.00)
    • Cached price up 400% ($0.0050 → $0.025)

Benchmark scores

Only one version of this line has a verifiable score set so far, so there is nothing to compare it against yet.

Which version to use

  • Top of the line: GPT-5.6 cyber

What your usage would cost

Adjust the monthly volume — every version of the line recomputes instantly.

GPT-5.6 cyber
$1250/mo
GPT-5.6 sol
$360/mo
GPT-5
$150/mo
GPT-5 mini
$30.00/mo
GPT-5 nano
$6.00/mo

List price only: cached input, batch discounts and off-peak rates are not included. The sample column prices 1M input + 500K output tokens per month.

Sources

Compare two versions of GPT-5

GPT-5.6 cyber vs GPT-5.6 solGPT-5.6 cyber vs GPT-5GPT-5.6 cyber vs GPT-5 miniGPT-5.6 cyber vs GPT-5 nanoGPT-5.6 sol vs GPT-5GPT-5.6 sol vs GPT-5 miniGPT-5.6 sol vs GPT-5 nanoGPT-5 vs GPT-5 miniGPT-5 vs GPT-5 nanoGPT-5 mini vs GPT-5 nano

All version series