Promo pricing is a loan, not a price, and two of the autumn's friendliest rates have expiry dates. Budget holders running high-volume workloads on either should model the post-promo numbers now.

Gemini 3.8 Flash: the January 1 doubling

Gemini 3.8 Flash (September 2) launched at a $0.75/$3.75 promo per million tokens that runs through December 31, 2026. On January 1, 2027 it doubles to $1.50/$7.50. Nothing about the model changes, only your bill, which goes up 100% on the stroke of midnight.

GPT-5.6 Sol promo: guaranteed through at least November 21

The GPT-5.6 family (July 9) carried promo pricing on Sol, the tier that became the ChatGPT default, and OpenAI's pricing page says that promotional pricing “is available at least through November 21, 2026.” That is a floor, not an end date: none has been announced. With GPT-6.1 Sol now listing at $2/$10 and GPT-6.1 Terra at $1/$5, the promo matters mainly to teams still pinned to GPT-5.6 model ids.

One counterweight: DeepSeek V4.1-Flash holds ~$0.15/$0.60 with MIT weights, and GPT-6 Luna sits at $0.10/$0.50. The budget tier is not disappearing, it is moving.

Our Reviews & Benchmarks table carries the per-model rates with promo flags, and we will update both entries on their respective change dates.

Update, October 27

Google's pricing page also dates a step-up for Gemini 4 Argon, from $2/$10 to $4/$20 on December 1; see our separate alert. This story originally said the GPT-5.6 Sol promo would end around November 21; OpenAI's wording is 'at least through', and we have corrected the text.