googlepricingllm-apideprecations

Gemini 2.5 Pro and 2.5 Flash Shut Down on December 1, 2026: What Replaces Them and What It Costs

Short answer. Google’s deprecations page, read on November 6, 2026, gives Gemini 2.5 Pro and Gemini 2.5 Flash a shutdown date of December 1, 2026, and Gemini Omni 1.1 Flash December 3. From those days requests route to Gemini 3.8 Flash, 3.5 Flash-Lite and Omni 1.2 Flash, billed at their rates. First seen by AI Signal on November 6, 2026.

In plain words. Two of Google’s older models stop answering on December 1. After that, Google sends the request to a newer model and charges that model’s price. For 2.5 Flash the price is the same. For 2.5 Pro it is cheaper until December 31, then dearer for input and cheaper for output.

What Google’s deprecations page says

Google’s deprecations page for the Gemini API has a table per model family with the columns “Model | Release date | Shutdown date | Recommended replacement |”. Read on November 6, 2026, three rows carry a date that was not there in the reading of October 26:

  • Under “Gemini 2.5 Pro models”: “gemini-2.5-pro | June 17, 2025 | December 1, 2026 | gemini-3.8-flash ( gemini-2.5-pro is no longer available in Google AI Studio; effective December 1, 2026, API requests to gemini-2.5-pro will automatically route to gemini-3.8-flash ) |”.
  • “gemini-2.5-flash | June 17, 2025 | December 1, 2026 | gemini-3.5-flash-lite ( gemini-2.5-flash is no longer available in Google AI Studio; effective December 1, 2026, API requests to gemini-2.5-flash will automatically route to gemini-3.5-flash-lite ) |”.
  • Under “Gemini Omni Flash models”: “gemini-omni-1.1-flash | August 27, 2026 | December 3, 2026 | gemini-omni-1.2-flash (Effective December 3, 2026, the endpoint is automatically routing API requests to gemini-omni-1.2-flash ) |”.

The note above the 2.5 Pro table adds: “Access through the API is limited to existing users who have actively used it in the past.” and “Review the breaking API changes in the Gemini 3.8 Flash migration guide before December 1, 2026.” The note above the 2.5 Flash table says the same for its migration guide, and settles the smaller model: “gemini-2.5-flash-lite is not deprecated and remains available through the API until further notice.”

Neither page gives the date of the announcement. AI Signal’s last reading of the deprecations page before this one, on October 26, 2026, had no shutdown date for any of the three; the dates were first seen on November 6, 2026. The deprecations page on this site lists them with the quotes above.

What happens after the date, in Google’s words

Google’s pricing page, read the same day, repeats the routing in each model’s block. For 2.5 Pro: “Gemini 2.5 Pro ( gemini-2.5-pro ) is deprecated and is no longer available to select in Google AI Studio. Effective December 1, 2026, all API requests to gemini-2.5-pro will automatically route to Gemini 3.8 Flash ( gemini-3.8-flash ) and be billed at Gemini 3.8 Flash rates”. For 2.5 Flash: “Gemini 2.5 Flash ( gemini-2.5-flash ) is deprecated and is no longer available to select in Google AI Studio. Effective December 1, 2026, all API requests to gemini-2.5-flash will automatically route to Gemini 3.5 Flash-Lite ( gemini-3.5-flash-lite ) and be billed at Gemini 3.5 Flash-Lite rates”.

So a call to the old model name after the date is not refused: it is answered by the replacement and billed at the replacement’s price. What that price is, and how it compares, is the rest of this post.

Gemini 2.5 Pro against Gemini 3.8 Flash: two comparisons

The replacement’s price has a date in it. Google’s pricing page, Standard tab, prints for Gemini 3.8 Flash: “Input price | Free of charge | $0.75 through December 31, 2026. $1.50 starting January 1, 2027. |”, “Output price (including thinking tokens) | Free of charge | $3.75 through December 31, 2026. $7.50 starting January 1, 2027. |” and “Context caching price | Free of charge | $0.075 through December 31, 2026. $0.15 starting January 1, 2027.” For Gemini 2.5 Pro the same tab prints “$1.25, prompts <= 200k tokens $2.50, prompts > 200k tokens |”, “$10.00, prompts <= 200k tokens $15.00, prompts > 200k |” and “$0.125, prompts <= 200k tokens $0.25, prompts > 200k”. Both as read on November 6, 2026; the dataset figures below were last verified on November 2.

a) Against the promotional price, through December 31, 2026.

Per million tokensGemini 2.5 ProGemini 3.8 Flash until Dec 31, 2026Bigger ÷ smaller
Input$1.25$0.751.25 ÷ 0.75 = 1.67×, replacement cheaper
Output$10.00$3.7510.00 ÷ 3.75 = 2.67×, replacement cheaper
Cached input$0.125$0.0750.125 ÷ 0.075 = 1.67×, replacement cheaper
Prompts over 200,000 tokens$2.50 in, $15.00 outone price for the whole window—

b) Against the list price, from January 1, 2027.

Per million tokensGemini 2.5 ProGemini 3.8 Flash from Jan 1, 2027Bigger ÷ smaller
Input$1.25$1.501.50 ÷ 1.25 = 1.20×, replacement dearer
Output$10.00$7.5010.00 ÷ 7.50 = 1.33×, replacement cheaper
Cached input$0.125$0.150.15 ÷ 0.125 = 1.20×, replacement dearer
Prompts over 200,000 tokens$2.50 in, $15.00 outone price for the whole window—

2.5 Pro’s page prints a second price for prompts over 200,000 tokens; 3.8 Flash’s page prints none, so a long prompt is billed at its single rate. The context window is the same in the dataset, 1,048,576 tokens for both.

The sample request. Take 10,000 input tokens and 1,000 output tokens, no cache, prompt under 200,000 tokens:

  • Gemini 2.5 Pro, today: 10,000 × $1.25 ÷ 1,000,000 = $0.0125, plus 1,000 × $10.00 ÷ 1,000,000 = $0.01, so $0.0225.
  • Gemini 3.8 Flash, December 1 to December 31, 2026: 10,000 × $0.75 ÷ 1,000,000 = $0.0075, plus 1,000 × $3.75 ÷ 1,000,000 = $0.00375, so $0.01125: 0.0225 ÷ 0.01125 = 2.00×, cheaper.
  • Gemini 3.8 Flash, from January 1, 2027: 10,000 × $1.50 ÷ 1,000,000 = $0.015, plus 1,000 × $7.50 ÷ 1,000,000 = $0.0075, so $0.0225: the same as 2.5 Pro today, to the cent, because the dearer input and the cheaper output cancel out at this ratio of input to output.

A request with more output than this one gets cheaper on the list price too; a request with less output gets dearer.

Gemini 2.5 Flash against Gemini 3.5 Flash-Lite: the same price

Google’s pricing page prints for Gemini 2.5 Flash “Input price | Free of charge | $0.30 (text / image / video) $1.00 (audio) |”, “Output price (including thinking tokens) | Free of charge | $2.50 |” and “Context caching price | Not available | $0.03 (text / image / video) $0.1 (audio) $1.00 / 1,000,000 tokens per hour (storage price) |”. For Gemini 3.5 Flash-Lite it prints “Input price | Free of charge | $0.30 (text / image / video / audio) |”, “Output price (including thinking tokens) | Free of charge | $2.50 |” and “Context caching price | Not available | $0.03 $1.00 / 1,000,000 tokens per hour (storage price) |”.

Per million tokensGemini 2.5 FlashGemini 3.5 Flash-LiteBigger ÷ smaller
Input (text)$0.30$0.301.00×, the same
Output$2.50$2.501.00×, the same
Cached input$0.03$0.031.00×, the same
Input (audio)$1.00$0.301.00 ÷ 0.30 = 3.33×, replacement cheaper

The sample request costs the same before and after: 10,000 × $0.30 ÷ 1,000,000 = $0.003, plus 1,000 × $2.50 ÷ 1,000,000 = $0.0025, so $0.0055 on both. The one line that moves is audio input, which the 2.5 Flash page prices apart at $1.00 and the 3.5 Flash-Lite page folds into the $0.30. Neither page prints a long-prompt tier for these two.

Gemini Omni 1.1 Flash against Omni 1.2 Flash

The third dated row is a video model. Google’s pricing page now heads the old row “Gemini Omni 1.1 Flash” and says: “Previous generation video generation and editing model. Effective December 3, 2026, all API requests to gemini-omni-1.1-flash automatically route to gemini-omni-1.2-flash and are billed at gemini-omni-1.2-flash rates.” Its Standard tab prints “$1.50 (text / image / video) |” for input and “$9.00 (text) $17.50 (video) * |” for output. The new row, Gemini Omni 1.2 Flash, released on November 5, 2026 according to the deprecations page, prints the same “$1.50 (text / image / video) |” for input and “$9.00 (text) $12.00 (video) * |” for output: text output at 1.00×, the same, and video output at 17.50 ÷ 12.00 = 1.46×, the new model cheaper.

Neither Omni row is in AI Signal’s main pricing table. The table covers models charged per token with text output, and Google’s model page describes Omni Flash as “a high-performance model for fast, conversational video generation and editing”; both rows are kept in the dataset, out of the table, on their own pages: Gemini Omni 1.1 Flash and Gemini Omni 1.2 Flash.

Where the two models sit until December 1

Both rows are now marked deprecated with their shutdown date in the Google API pricing page on this site and in the pricing table, where deprecated rows are shown behind a toggle. Their prices stay as last verified on November 2, 2026, and the weekly check keeps reading them on Google’s pricing page until the date. After December 1 the rows become retired, with the price they had, and the page of each keeps saying what replaced it.

How we checked

  • Deprecations: Google’s Gemini API deprecations page, aistudio.google.com/docs/deprecations, read by script on November 6, 2026, table rows quoted as printed (“Model | Release date | Shutdown date | Recommended replacement |”). The previous reading, on October 26, 2026, had no shutdown date for these three models.
  • Prices: Google’s pricing page, aistudio.google.com/docs/pricing, Standard tab of each model’s block, read by script on November 6, 2026. The dataset rows (Gemini 2.5 Pro, 2.5 Flash, 3.8 Flash, 3.5 Flash-Lite) were last verified on November 2 and matched the page on November 6.
  • The move of Google’s documentation: since at least November 3, 2026, every page under ai.google.dev/gemini-api/docs/ answers with a 301 redirect to aistudio.google.com/docs/, and the new pages are drawn in the browser. This site reads them from the Markdown version Google serves at the same address with “.md”; the quotes above are from that text.
  • Arithmetic: the sample request and every ratio are computed here from the prices quoted, and rounded only in the text.
  • Omni: Google’s model page for both Omni Flash versions, aistudio.google.com/docs/models/gemini-omni-flash, read on November 6, 2026, for the description quoted; the pricing page for the figures.

Frequently asked questions

When will Gemini 2.5 Pro be shut down?

On December 1, 2026, according to Google's deprecations page read on November 6, 2026, which lists 'gemini-2.5-pro | June 17, 2025 | December 1, 2026 | gemini-3.8-flash'. Google's pricing page says that from that day requests to gemini-2.5-pro 'will automatically route to Gemini 3.8 Flash ( gemini-3.8-flash ) and be billed at Gemini 3.8 Flash rates'.

What replaces Gemini 2.5 Flash?

Gemini 3.5 Flash-Lite: the deprecations page lists 'gemini-2.5-flash | June 17, 2025 | December 1, 2026 | gemini-3.5-flash-lite', and says requests 'will automatically route to gemini-3.5-flash-lite and be billed at Gemini 3.5 Flash-Lite rates'. Both cost $0.30 per million input tokens and $2.50 per million output tokens in AI Signal's dataset, with cached input at $0.03.

Is the replacement cheaper or more expensive?

For 2.5 Pro it depends on the date: Gemini 3.8 Flash is $0.75 in and $3.75 out through December 31, 2026 (1.67 and 2.67 times cheaper than 2.5 Pro's $1.25 and $10.00), and $1.50 and $7.50 from January 1, 2027 (input 1.20 times dearer, output 1.33 times cheaper). For 2.5 Flash the price does not move: 3.5 Flash-Lite costs the same $0.30 and $2.50.

Does Gemini 2.5 Flash-Lite shut down too?

No: the deprecations page, read on November 6, 2026, says 'gemini-2.5-flash-lite is not deprecated and remains available through the API until further notice'. Its model page says 'gemini-2.5-flash-lite is not deprecated and will continue to be served until further notice through the API'.

Drafted with AI assistance from the cited source and reviewed by Álvaro Lucero Rufino, who is responsible for what stays published here. Automated publishing was discontinued in September 2026; posts that no longer met that standard were removed rather than left up. This is original explanatory writing, not a reproduction — read the source above for the full story.