Every scheduled 2027 API price increase, in one table
Three Gemini Flash models and both Gemini TTS models double on 1 January 2027, while Anthropic cancelled a planned rise. The full list, with what it costs you.
Model prices usually fall. Three of Google's Gemini Flash models and both of its text-to-speech models are scheduled to double on 1 January 2027, and the increase is printed on Google's own pricing page today. If your budget assumes the current rate, it is half of what you will pay from January.
What rises on 1 January 2027
| Model | Now, per 1M tokens | From 1 January 2027 |
|---|---|---|
| Gemini 3.8 Flash | $0.75 in / $3.75 out | $1.50 in / $7.50 out |
| Gemini 3.7 Flash | $0.75 in / $3.75 out | $1.50 in / $7.50 out |
| Gemini 3.6 Flash | $0.75 in / $3.75 out | $1.50 in / $7.50 out |
| Context caching, all three | $0.075 | $0.15 |
| Cache storage, all three | $0.50 per 1M tokens per hour | $1.00 |
| Gemini 3.8 Flash TTS | $0.50 text in / $9.00 audio out | Doubles |
| Gemini 3.8 Flash-Lite TTS | $0.50 text in / $6.00 audio out | Doubles |
Batch processing stays at half the standard price, so batch users see the same doubling from a lower base. The Flash-Lite text models, Gemini 3.5 Flash-Lite and 3.1 Flash-Lite, are not on the list and keep their current prices.
Who is affected, and by how much
A service that sends 10 million input tokens and 2 million output tokens a month on Gemini 3.8 Flash pays $15.00 today. The same month costs $30.00 from January.
The models affected are Google's mid-tier Flash line, which is where most production traffic sits. Three alternatives cost less than the 2027 Flash price today: Gemini 3.5 Flash-Lite at $0.30 and $2.50, GPT-6 Luna at $0.10 and $0.50, and DeepSeek V4.1 Flash at $0.15 and $0.60 off-peak. Whether any of them does your job is a separate question, and a measurable one.
Watch for "introductory" and "limited-time" wording elsewhere. It is becoming common, and it means today's number has an expiry date that your budget probably does not.
Prices moving the other way
- Anthropic cancelled an increase. Claude Sonnet 5's launch price of $2 and $10 per million was announced as introductory, with a rise to $3 and $15 scheduled for 1 September 2026. Anthropic's pricing page now states the increase will not happen and $2 / $10 is the standard price.
- Anthropic cut the Opus tier. Claude Opus 5.5 lists $4 and $20 against $5 and $25 for Opus 5, with cache reads down from a tenth of the input price to a twentieth.
- OpenAI halved two tiers. GPT-6 Sol costs $2 and $10 where GPT-5.6 Sol cost $4 and $20. GPT-6 Luna costs $0.10 and $0.50 where GPT-5.6 Luna cost $0.20 and $1.20.
- DeepSeek split its prices by time of day. Since 10 September its rates double between 01:00 and 04:00 and between 06:00 and 10:00 UTC on weekdays.
The overall direction is still downward at the top of the market and upward in Google's mid tier. That makes a single "what does AI cost" number useless, and a per-workload measurement necessary.
What to do before January
- Re-price any Gemini Flash workload at the 2027 rate. If it still works at $1.50 and $7.50, nothing to do.
- Test one tier down. Gemini 3.5 Flash-Lite costs less than half the 2027 Flash price and keeps the same image and document handling.
- Test one lab across. GPT-6 Luna and DeepSeek V4.1 Flash are both below the current Flash price, let alone the 2027 one.
- Check for other expiry dates. Any price described as introductory or limited-time should carry a reminder in your own notes.
Price your own prompt on all four with the token calculator, or start from the comparison hub.
Evidence ledger
Reviewed: 2 October 2026 · Provider pricing pages remain authoritative and change without notice.
- Gemini API pricing, which lists both the current and the 2027 rates
- Anthropic pricing, including the cancelled Sonnet 5 increase
- OpenAI API pricing
- DeepSeek models and pricing