Gemini 3.1 Pro vs GPT-5.6 Terra: API Cost Compared
The verdict, in numbers: Gemini 3.1 Pro and GPT-5.6 Terra publish identical list rates ($2.00 / $12.00), but identical list rates do not mean identical bills: on our 200-step agent workload the two land 1.0x apart ($2.79 vs $2.89), driven entirely by the cache-read row.
This page compares list rates and workload cost only. It does not tell you which model is better — capability belongs to your own evaluation. What it does give you is the half of the decision that can be verified: what each provider will actually charge, read from their own documentation on September 18, 2026.
Rate cards side by side
| Item | Gemini 3.1 Pro | GPT-5.6 Terra |
|---|---|---|
| Provider | Google DeepMind | OpenAI |
| Input / 1M | $2.00 | $2.00 |
| Output / 1M | $12.00 | $12.00 |
| Cache read / 1M | $0.20 | $0.20 |
| Context window | 1M | 1.05M |
Sources: Google DeepMind pricing ↗ and OpenAI pricing ↗, both checked September 18, 2026.
The same 200-step agent task, priced twice
One coding-agent task: 200 steps, a stable 40,000-token prefix written to cache on step 1 and re-read on the remaining 199, 500 output tokens per step. Cache-write billing follows each provider’s own rules (cache writes bill at the input rate (per-hour cache storage is excluded); cache writes bill at 1.25x the uncached input rate).
| Component | Gemini 3.1 Pro | GPT-5.6 Terra |
|---|---|---|
| Cache write (once) | — | $0.10 |
| Cache reads (199 × 40K) | $1.59 | $1.59 |
| Output (200 × 500) | $1.20 | $1.20 |
| Total per task | $2.79 | $2.89 |
A heavy month on each
50M fresh input, 150M cached input, 10M output tokens in a month:
| Component | Gemini 3.1 Pro | GPT-5.6 Terra |
|---|---|---|
| Fresh input (50M) | $100.00 | $100.00 |
| Cached input (150M) | $30.00 | $30.00 |
| Output (10M) | $120.00 | $120.00 |
| Monthly total | $250.00 | $250.00 |
What this comparison does and does not say
Gemini 3.1 Pro: Google flagship for reasoning, agentic work, and long context; multimodal input.
GPT-5.6 Terra: Balanced GPT-5.6 tier for strong capability at a lower unit price.
Cost is the verifiable half of the decision; quality is the half only your workload can answer. Run both candidates on a representative evaluation set and measure cost per accepted task — the cost planning guide walks through that method, and the calculator lets you re-price this exact comparison with your own token mix.
Frequently asked questions
Is Gemini 3.1 Pro cheaper than GPT-5.6 Terra?
On list rates, Gemini 3.1 Pro costs $2.00/$12.00 per 1M input/output tokens versus $2.00/$12.00 for GPT-5.6 Terra. On a 200-step agent workload (40K cached prefix, 500 output tokens per step), Gemini 3.1 Pro totals $2.79 versus $2.89 for GPT-5.6 Terra. Prices verified September 18, 2026.
Gemini 3.1 Pro vs GPT-5.6 Terra: which has the longer context window?
Gemini 3.1 Pro offers 1M tokens; GPT-5.6 Terra offers 1.05M tokens, per provider documentation checked September 18, 2026.
Where do these Gemini 3.1 Pro and GPT-5.6 Terra prices come from?
Both rate cards were read from the providers' own pricing pages on September 18, 2026: https://deepmind.google/models/gemini/pro/ and https://developers.openai.com/api/docs/models/gpt-5.6-terra.