Coding & Dev 📅 2026-11-19 · Mini AI Tools Research Desk

gemini-2.5-flash Is Off the Picker: Dec 1 Sends It to Flash-Lite

On 19 Nov 2026, gemini-2.5-flash is off the Studio picker. From 1 Dec calls route to gemini-3.5-flash-lite. The quickstart sample is gemini-4-hydrogen.

$ EXECUTIVE BENCHMARK SUMMARY (TL;DR)

Two old ids leave the picker on 1 December

On 19 November 2026 the Gemini pricing page says gemini-2.5-flash is deprecated and is no longer available to select in Google AI Studio. Effective 1 December 2026, all API requests to that id automatically route to gemini-3.5-flash-lite and are billed at Gemini 3.5 Flash-Lite rates.

The same page says gemini-2.5-pro is deprecated and is off the picker. Effective 1 December 2026, requests to that id route to gemini-3.8-flash and are billed at Gemini 3.8 Flash rates. On 1 December 2026, a body that still names the old id is routed and billed as the destination model. No generate call was sent for this page. Those two clocks are the deprecation lines on the pricing page.

gemini-2.5-flash routes to Flash-Lite on 1 Dec, gemini-2.5-pro routes to 3.8 Flash, and the quickstart sample is gemini-4-hydrogen🔍 View full resolution
Old ids, their 1 Dec destinations, and the id in the getting-started sample. 19 Nov 2026.

Pricing: aistudio.google.com/docs/pricing

The getting-started sample is gemini-4-hydrogen

The getting-started page the same day uses the Interactions API. Step 1 says Google AI Studio creates a project and an API key for a new user, and that the key is copied from the API keys page. Additional projects, and projects imported from Google Cloud, may need project verification before they can use the free tier. The page says to raise rate limits with Set up billing on the API keys page. It does not print a requests-per-minute number.

The rate-limits page says live limits are viewed in AI Studio, that specified limits are not guaranteed, and that a second key in the same project does not add a second quota. Daily quotas reset at midnight Pacific time. The only Hydrogen figure on that page is in the batch table: 3,000,000 enqueued tokens. Treat the 3,000,000 as a batch queue. The interactive cap for interactions.create is a different number, and this page does not print it. The worked example "an RPM limit of 20" shows how any cap is counted. The page does not attach that 20 to Hydrogen.

Step 2 installs google-genai and calls client.interactions.create with model="gemini-4-hydrogen" and a short input string. genai.Client() reads GEMINI_API_KEY.

Getting started page: a new user gets a project and API key, then step 2 begins🔍 View full resolution
Step 1 on the getting-started page: a new user gets a project and a key. The sample model is written in step 2. 19 Nov 2026.

Get a key: aistudio.google.com/apikey

Try it: aistudio.google.com/docs/get-started

Rate limits: aistudio.google.com/docs/rate-limits

Free-tier cells, including the search cell

Hydrogen, the quickstart id: free-tier input, output, and context caching say Free of charge. Paid tier is $1.50 input and $7.50 output per 1 million tokens. Grounding with Google Search on the free tier says Not available. Used to improve our products is Yes on the free tier and No on the paid tier.

Flash-Lite, the 1 December destination for gemini-2.5-flash: free-tier input and output say Free of charge. Paid input is $0.30 per 1 million tokens for text, image, video, and audio. Paid output is $2.50. Free-tier Google Search says Not available. The old gemini-2.5-flash row still prints free Google Search up to 500 requests a day, shared on the page with gemini-2.5-flash-lite. On 1 December 2026 the page routes gemini-2.5-flash. It does not say gemini-2.5-flash-lite is routed. A call that still needs those 500 free searches uses gemini-2.5-flash-lite. The destination row, gemini-3.5-flash-lite, does not list that cell.

3.8 Flash, the destination for gemini-2.5-pro: free-tier input and output say Free of charge. Paid input is $0.75 per 1 million through 31 December 2026 and $1.50 from 1 January 2027. Paid output is $3.75, then $7.50. Free-tier Google Search says Not available.

Carbon, Argon, and Neon: the free-tier input cell and the free-tier output cell both say Not available. Carbon's paid input is $2.00 per 1 million through 30 November 2026 and $4.00 from 1 December 2026. Paid output is $10.00, then $20.00. Text rates on the old gemini-2.5-flash paid row and the destination gemini-3.5-flash-lite paid row both print $0.30 input and $2.50 output. Audio input on the old row is $1.00. The gemini-3.5-flash-lite input cell lists $0.30 for audio as well. The pricing page says the rerouted calls are billed at Flash-Lite rates. From 1 December 2026 the bill uses the destination row: gemini-3.5-flash-lite for a gemini-2.5-flash call, and gemini-3.8-flash for a gemini-2.5-pro call. The old row is the price printed for that id before the route.

Free-tier cells for Hydrogen, Flash-Lite, 3.8 Flash, and Carbon🔍 View full resolution
Free-tier cells and the paid prices printed beside them. 19 Nov 2026.
Model idWhat happens to a callFree tier input and outputPaid price per 1M, as printed
gemini-2.5-flashOff the picker. 1 Dec routes to gemini-3.5-flash-liteFree of charge. Search: up to 500 RPD, shared with gemini-2.5-flash-liteInput $0.30 text/image/video, $1.00 audio. Output $2.50
gemini-3.5-flash-liteDestination of the 2.5 Flash routeFree of charge. Search: Not availableInput $0.30 for text, image, video, and audio. Output $2.50
gemini-2.5-proOff the picker. 1 Dec routes to gemini-3.8-flashFree of charge. Search: Not availableInput $1.25 at or under 200k prompt tokens, $2.50 above. Output $10 at or under 200k, $15 above
gemini-3.8-flashDestination of the 2.5 Pro routeFree of charge. Search: Not availableInput $0.75 through 31 Dec 2026, then $1.50. Output $3.75, then $7.50
gemini-4-hydrogenThe id in the getting-started sampleFree of charge, caching included. Search: Not availableInput $1.50. Output $7.50
gemini-4-carbonFree-tier input and output both say Not availableNot availableInput $2.00 through 30 Nov 2026, then $4.00. Output $10, then $20

Pricing: aistudio.google.com/docs/pricing

Which string to put in the call

Choose gemini-4-hydrogen if you are copying the first working sample on the getting-started page and the key has no billing account. Free-tier input, output, and caching say Free of charge. That free-tier row does not include Google Search. Prompts on the free tier can be used to improve Google products. The paid row says No.

Get a key: aistudio.google.com/apikey

If the code still says gemini-2.5-flash and the log should show the id you typed, change it to gemini-3.5-flash-lite now. On 1 December 2026 the pricing page routes that id. Those 500 free searches a day stay with gemini-2.5-flash-lite, the id the page does not say it will route. The destination row says Not available.

If the code still says gemini-2.5-pro, the destination is gemini-3.8-flash. Free-tier input and output say Free of charge. Paid input stays $0.75 per 1 million tokens through 31 December 2026, then $1.50 from 1 January 2027.

Try it: aistudio.google.com/docs/get-started

Carbon, Argon, and Neon say Not available on both the free-tier input cell and the free-tier output cell. Leave those three ids out of a request until billing is set up. On the getting-started sample, the id that still has a free input cell is hydrogen.

Pricing: aistudio.google.com/docs/pricing

Rate limits: aistudio.google.com/docs/rate-limits

Featured Tools in this Guide

GR

Groq

Coding & Dev
Freemium

High-speed LLM inference API powered by LPU hardware, offering free developer quotas for open-weight and speech models.

GA

Gemini API

Coding & Dev
Freemium

Google's official multimodal developer API, offering free-tier prototyping on Gemini 4 Hydrogen and Flash models.

OP

OpenRouter

Coding & Dev
Freemium

Unified AI model router providing single-API access to hundreds of commercial and open-weight models, including free tiers.

GC

Google Colab

Productivity & Auto
Freemium

Google's hosted Jupyter notebooks with one-click GPU, Drive integration and zero setup.

More In-Depth Benchmarks & Guides

View all guides →