AIVX Labs
AI news for agencies · Models and costs

Gemini 3.8 Flash's intro price ends December 31: bulk jobs to run now

By Alan Magliocca · Published November 2026

Quick answer: Google's introductory API price for Gemini 3.8 Flash, $0.75 per million input tokens and $3.75 per million output tokens, runs through December 31, 2026, and doubles to $1.50 and $7.50 starting January 1, 2027. If you have bulk text jobs for clients, measure them on a small sample now and run the full batches before December 31.

Key takeaways

  • Gemini 3.8 Flash costs $0.75 input and $3.75 output per million tokens through December 31, 2026.
  • From January 1, 2027, the standard price is $1.50 input and $7.50 output per million tokens.
  • Batch and Flex cost half the standard price, both before and after the change.
  • Output pricing includes thinking tokens, and higher effort levels use more of them.
  • Run a 20-item sample first to measure real token use, then schedule the full job before December 31.

What happened

On September 2, 2026, Google released Gemini 3.8 Flash. Google calls it its most intelligent workhorse model, with gains over Gemini 3.7 Flash in software engineering, agentic tasks and multi-step reasoning, at the speed and low cost of 3.7 Flash.

The launch price is introductory. Google's announcement says it "expires on December 31, 2026." Google's Gemini API pricing page lists the paid prices per million tokens:

  • Standard: $0.75 input and $3.75 output through December 31, 2026. From January 1, 2027: $1.50 input and $7.50 output.
  • Batch and Flex: $0.375 input and $1.875 output through December 31, 2026. From January 1, 2027: $0.75 input and $3.75 output.
  • Context caching: $0.075 through December 31, 2026, then $0.15 from January 1, 2027.

The output price includes thinking tokens. The pricing page also lists a free tier for standard use.

Where you can use it: developers through the Gemini API in Google AI Studio, Google Antigravity, Android Studio and Stitch; businesses through Gemini Enterprise; and Google AI Pro and Ultra subscribers in the Gemini app, AI Mode in Search and Gemini in Google Sheets.

Google adds one caution. The model does more work on complex tasks, with extra reasoning steps and repeated tool calls, so it can use more tokens, especially at higher effort levels. Google says you can use lower effort levels to cut token use, or keep using Gemini 3.7 Flash, which remains supported.

What it means for agency owners

Any bulk text job you were planning for January costs half as much if it runs by December 31. That covers the kind of work agencies do in volume for clients: product and service descriptions, meta descriptions, tagging reviews, cleaning up CRM exports and drafting FAQ answers.

The per-job dollar amounts are small, but they add up across many clients and repeat runs. If you've quoted clients a fixed monthly price that includes Gemini 3.8 Flash usage, the cost of that work doubles on January 1, 2027. Check your margins now.

The effort setting matters as much as the price. Google says the model can use more tokens at higher effort levels, and thinking tokens are billed as output. Measuring a sample tells you what a job really costs.

What to do this week

  1. List the bulk jobs for each client that don't depend on new information: product descriptions, meta descriptions, review tagging, FAQ drafts and CRM category cleanup.
  2. Run a 20-item sample at low effort. Read the token counts from your API usage, then raise the effort only if the quality falls short.
  3. Use Batch for anything that isn't urgent. It costs half the standard price before and after the change.
  4. Schedule the full runs well before December 31, 2026, and leave time for a person to review the output before it goes on a client's site.
  5. Reprice 2027 work. Update any fixed-price client quotes that include Gemini 3.8 Flash usage at the January 1 rate.

How AIVX Labs helps

If you'd rather not manage API keys, token counts and price changes per model, AIVX Labs tools run on credits instead. The AI Content Studio, email sequences, landing page sales copy and the Review Analyzer cover much of the bulk writing and sorting agencies do, and the Prospect Finder returns 25 local businesses per search for 2 credits. Tell RICO the client job in plain words and it picks the tools, shows the cost in credits and runs only when you click Go. For larger jobs it proposes a plan of 2 to 6 tools with one total cost to approve.

What does a bulk job cost before and after December 31?

Take an e-commerce client with 2,000 product descriptions to rewrite. Assume each request sends about 1,000 tokens of instructions and product data, and returns about 300 tokens of text. That's 2 million input tokens and 0.6 million output tokens.

Standard, through December 31, 2026:

  • Input: 2 million × $0.75 = $1.50
  • Output: 0.6 million × $3.75 = $2.25
  • Total: $3.75

Standard, from January 1, 2027:

  • Input: 2 million × $1.50 = $3.00
  • Output: 0.6 million × $7.50 = $4.50
  • Total: $7.50

Batch, through December 31, 2026: $0.75 + $1.125 = about $1.88. Batch from January 1, 2027: $3.75.

Two caveats. First, thinking tokens count as output, so at higher effort the real output count can be well above 300 tokens per item; your sample run shows the true number. Second, the model cost is the small part. A person reviewing 2,000 descriptions costs far more than the tokens, so build review time into your quote.

Is Gemini 3.8 Flash the cheapest option for every bulk job?

Not always. Match the model to the job.

  • Simple sorting and tagging: a small model may be enough. Anthropic lists Claude Haiku 5.5 at $0.10 input and $0.50 output per million tokens for prompts up to 100k tokens; see Haiku 5.5 at 10 cents per million tokens.
  • Writing and multi-step reasoning at volume: Gemini 3.8 Flash is built for this, and its introductory price makes it cheap through December 31.
  • Hard, long jobs where quality matters most: a frontier model may earn its price. Google's Gemini 4 Argon has its own introductory price ending sooner, on November 30; see Gemini 4 Argon now powers AI Mode, and its API price doubles November 30. For how Claude's top model compares, read Claude Opus 5.5 cut prices about 20%.

A practical rule: run the same 20-item sample on two models and compare cost per usable item, not cost per token. A cheaper model that needs more rewrites isn't cheaper.

One example from the webinar side: if a client runs webinars, a bulk job could be first drafts of reminder and follow-up emails for every event. BrightStage's guide to webinar email sequences covers how those emails are usually structured.

Sources

Written by Alan Magliocca, founder of AIVX Labs and BrightStage AI (Podavinci LLC). Written with AI assistance and checked against the linked sources.

Read our featured article on LinkedIn

Get AIVX Labs verified for $1: RICO, your AI operator, with 7 days of full access.

Get AIVX Labs Verified for $1

Follow along:LinkedInYouTubeTikTokXFacebookInstagramSkool community

Frequently asked questions

Will a batch job submitted on December 31 get the introductory price?

Google's pricing page lists the dates but doesn't say how jobs that span the change are billed. Finish large jobs a few days early to be safe.

Does the price change affect Gemini app subscribers?

No. These are Gemini API prices. Google AI Pro and Ultra subscribers use Gemini 3.8 Flash in the Gemini app, AI Mode and Sheets as part of their subscription.

Will Gemini 3.7 Flash stay available?

Google says Gemini 3.7 Flash remains supported, so you can keep using it for jobs where 3.8 Flash's extra reasoning isn't needed.