Checked against the vendor's own page, and dated.

GPT-4o shuts off 14 Jan 2027: what its named replacement costs at three monthly volumes

OpenAI will remove GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini from its API on 14 Jan 2027 and names gpt-6-luna as the replacement. Tooldossier read OpenAI's API pricing and deprecations pages on 24 Oct 2026.

Model Input per 1M tokens Cached input per 1M Output per 1M Status Checked
GPT-4o $2.50 $1.25 $10.00 removed 14 Jan 2027 24 Oct 2026
GPT-4o mini $0.15 $0.075 $0.60 removed 14 Jan 2027 24 Oct 2026
GPT-4.1 $2.00 $0.50 $8.00 removed 14 Jan 2027 24 Oct 2026
GPT-4.1 mini $0.40 $0.10 $1.60 removed 14 Jan 2027 24 Oct 2026
GPT-6 Luna $0.10 $0.01 $0.50 named replacement 24 Oct 2026
GPT-6.1 Sol $2.00 $0.10 $10.00 current flagship 24 Oct 2026

All prices are OpenAI's Standard processing rates at short context, per 1 million tokens, from the OpenAI API price file. There is no per-seat fee; the bill is tokens used.

What OpenAI is retiring

OpenAI's deprecations page says developers using certain GPT-4 series models were notified on 14 Oct 2026 that the models will be removed from the API on 14 Jan 2027:

Model API snapshots Shutdown date Recommended replacement
GPT-4o gpt-4o-2024-08-06, gpt-4o-2024-11-20 14 Jan 2027 gpt-6-luna
GPT-4o mini gpt-4o-mini-2024-07-18 14 Jan 2027 gpt-6-luna
GPT-4.1 gpt-4.1-2025-04-14 14 Jan 2027 gpt-6-luna
GPT-4.1 mini gpt-4.1-mini-2025-04-14 14 Jan 2027 gpt-6-luna

OpenAI names one replacement for all four. The pricing page describes GPT-6 Luna as its most efficient model for focused, high-volume tasks, and GPT-6.1 Sol as built for complex coding and agentic workflows. Tooldossier does not judge output quality; the sums below are price only.

What the switch costs per month

Three monthly volumes, each with five input tokens for every output token and no caching, at the prices above:

10 million input and 2 million output tokens a month

  • GPT-4o: 10 x $2.50 + 2 x $10.00 = $45
  • GPT-4.1: 10 x $2.00 + 2 x $8.00 = $36
  • GPT-4.1 mini: 10 x $0.40 + 2 x $1.60 = $7.20
  • GPT-4o mini: 10 x $0.15 + 2 x $0.60 = $2.70
  • GPT-6 Luna: 10 x $0.10 + 2 x $0.50 = $2
  • GPT-6.1 Sol: 10 x $2.00 + 2 x $10.00 = $40

100 million input and 20 million output tokens a month

  • GPT-4o: 100 x $2.50 + 20 x $10.00 = $450, against GPT-6 Luna at 100 x $0.10 + 20 x $0.50 = $20 and GPT-6.1 Sol at 100 x $2.00 + 20 x $10.00 = $400
  • GPT-4o mini: 100 x $0.15 + 20 x $0.60 = $27, against GPT-6 Luna at $20

1 billion input and 200 million output tokens a month

  • GPT-4o: 1,000 x $2.50 + 200 x $10.00 = $4,500, against GPT-6 Luna at 1,000 x $0.10 + 200 x $0.50 = $200
  • GPT-4.1: 1,000 x $2.00 + 200 x $8.00 = $3,600, against GPT-6.1 Sol at 1,000 x $2.00 + 200 x $10.00 = $4,000

Who pays more, who pays less

On these prices, moving to the named replacement lowers the token bill for all four retiring models. GPT-4o to GPT-6 Luna at the first volume is ($45 - $2) / $45, or 96% less. GPT-4o mini to GPT-6 Luna is ($2.70 - $2) / $2.70, or 26% less.

A team that moves to GPT-6.1 Sol instead pays less than on GPT-4o ($40 against $45 at the first volume) but more than on GPT-4.1 ($40 against $36), because Sol's output price, $10.00, is above GPT-4.1's $8.00.

Prices outside these sums are on the file: cached input, which lowers the bill for repeated prompts, and a 10% uplift on regional processing endpoints for models released on or after 5 Mar 2026. Batch, Fast and long context prices are listed separately on OpenAI's page and are not used here.

More files will be added to the AI model APIs drawer.

Every price above was read from each vendor's own pricing page on the date shown, and each file links the captured page.

Every figure in this note is taken from a dated price file on this site and links to it; each file links to the vendor page the price was read from. Spotted an error? How corrections work.