When does Gemini 4 Argon's price increase?

December 1, 2026. Google's own pricing docs list Gemini 4 Argon at $2.00 input / $10.00 output per 1M tokens 'through November 30, 2026' and $4.00 / $20.00 'starting December 1, 2026' -- a flat doubling of input, output, batch, and context-caching rates on that date.

Is Gemini 3.8 Flash the same as Gemini 3.7 Flash?

Gemini 3.8 Flash is the replacement for Gemini 3.7 Flash, which Google deprecated on September 2, 2026. API requests to the old gemini-3.7-flash model ID are automatically routed to gemini-3.8-flash, and 3.7 is no longer selectable in Google AI Studio. The introductory pricing carried over unchanged: $0.75 input / $3.75 output per 1M tokens through December 31, 2026, rising to $1.50 / $7.50 on January 1, 2027.

Why do Gemini's promotional prices expire on different dates?

Because each model's introductory window is tied to its own launch date, not a shared calendar deadline. Gemini 4 Argon's discount ends November 30, 2026. The Gemini 3.8 Flash family -- plus Flash TTS, Flash-Lite TTS, and the Robotics ER 2 Preview models -- end a month later, December 31, 2026, with new rates from January 1, 2027. Budgeting against 'current Gemini pricing' as one number ignores that it is really two numbers with two expiration dates.

Did this change what your AI cost calculator shows for Gemini?

Yes. Before today, neither Gemini 4 Argon nor Gemini 3.8 Flash appeared anywhere in our model list, even though the newest flagship models from OpenAI and Anthropic were already there. Both are now added at their current, pre-cliff prices. We'll need to update them again on December 1 and January 1 when the promotional rates expire -- the same habit we're applying to every provider, not a one-time fix.

Gemini 4 Argon Is $2 Today. It's $4 on December 1. We Didn't Have It in Our Calculator Yet.

2 November 2026 · Site Methodology & Pricing Notes · 5 min read

Part of keeping this site honest is re-checking provider docs on a normal day, not just when something looks obviously wrong. Today's normal check on Google's Gemini pricing page turned up two things worth a post: a real price cliff 29 days away that nothing on this site mentioned, and a model name on our own Gemini page that Google itself retired two months ago.

Gemini 4 Argon's intro price doubles December 1

Straight from Google's current Gemini API pricing docs, checked 2026-11-02: Gemini 4 Argon -- the new flagship, with a unified 1M-token context window -- is priced at $2.00 input / $10.00 output per 1M tokens "through November 30, 2026," rising to $4.00 / $20.00 "starting December 1, 2026." Batch pricing doubles the same day, from $1.00/$5.00 to $2.00/$10.00. Context caching doubles too, $0.10 to $0.20.

Avaliar	Through Nov 30, 2026	From Dec 1, 2026
Entrada / 1 milhão de tokens	$2.00	$4.00
Saída/1 milhão de tokens	$10.00	$20.00
Batch input / output	$1.00 / $5.00	$2.00 / $10.00
Context caching / 1M	$0.10	$0.20

That's a flat 2x across every rate on the page, on a single named date. Nothing in the docs frames it as a "promotional period ending soon" the way some providers hedge it -- it's just two columns, before and after.

It isn't the only Gemini cliff -- and it isn't the same date

Gemini 3.8 Flash -- the current Flash workhorse, positioned for "long-horizon software engineering, autonomous agents, and complex enterprise workflows" -- has its own introductory rate: $0.75 input / $3.75 output through December 31, 2026, doubling to $1.50 / $7.50 from January 1, 2027. Same pattern Google used for Gemini 3.8 Flash TTS and the Robotics ER 2 Preview models. All of them land on January 1, not December 1.

So "current Gemini pricing" isn't one number with one expiration date. It's at least two numbers, 31 days apart: the flagship's discount ends with November, the whole Flash family's discount ends with the year. Anyone budgeting a Q1 2027 AI spend against "Gemini's current rates" needs both dates, not one.

Neither new model was in our calculator

Here's the part that isn't Google's fault. Checking those numbers against our own Calculadora de custos de IA's model list, Gemini 4 Argon and Gemini 3.8 Flash were both simply absent -- while GPT-6 Astra and Claude Opus 5.5, the equivalent newest flagships from OpenAI and Anthropic, were already there. Google's two newest models had a real, dated price change coming and weren't in the comparison at all. Both are added now, at today's pre-cliff rates.

And our own Gemini page was naming a dead model

Second finding, smaller but more embarrassing: our Gemini pricing page listed Gemini 3.7 Flash in its pricing table, in all eight languages. Google deprecated Gemini 3.7 Flash on September 2, 2026 -- two months before today's check. API requests to the old model ID are silently auto-routed to gemini-3.8-flash, and 3.7 isn't even selectable in AI Studio anymore.

The dollar figures on our page were already correct, because the introductory rate carried over unchanged in the swap. The name just hadn't been updated since whatever August snapshot the page was last built from. Fixed across all eight locale files today.

What I'm changing

Added Gemini 4 Argon and Gemini 3.8 Flash to the live model list behind the Calculadora de custos de IA e o Calculadora de custos da API Gemini, and corrected the stale "3.7 Flash" name on the Gemini vendor page in all eight languages. The bigger habit: a brand-new model name is itself a signal to re-check, not just a price. Google didn't just reprice Gemini 3.7 -- it retired the name entirely while keeping the number. A page that only tracks dollar figures can stay "accidentally correct" on price while being two model generations out of date on name, and nothing about that mismatch is visible unless someone goes back to the source docs looking for it.

Gemini API pricing & free tier — model name corrected to 3.8 Flash in all 8 languages.
Calculadora de custos da API Gemini — now includes Gemini 4 Argon and Gemini 3.8 Flash.
Full AI API cost calculator — compare Argon and 3.8 Flash against every other provider's newest models.

New to LLM pricing in general? Start with the guias gratuitos de custo de API ou o completo AI & LLM API directory.

Implemente você mesmo: DigitalOcean – crédito grátis de $ 200 ↗ · VPS Hostinger ↗

Checked directly against ai.google.dev/gemini-api/docs/pricing and ai.google.dev/gemini-api/docs/models, 2026-11-02. Reference figures, not live quotes -- confirm current terms on Google's own docs before relying on a cutoff date for production budgeting.