What Gemini 3.1 Pro Is
Gemini 3.1 Pro is Google's frontier reasoning model — leading on novel logic (ARC-AGI-2 77.1%), graduate-level science (GPQA Diamond 94.3%), and agentic reliability, with a 1M-token context (up to 2M) and full multimodal input. It's built for the hardest software engineering, long-document analysis, and autonomous agent workflows.
Pricing
Per 1M tokens: ~$2 input / ~$12 output for prompts ≤200K; above 200K it steps up to ~$4 / ~$18. Batch API halves it (~$1/$6); context caching cuts cached input ~90%. Note the 2.5 family bills thinking tokens in output, and Pro tiers moved to paid-only in April 2026. On LinkModel it runs behind one key with Claude, GPT, and DeepSeek, up to 30% below official.
| ≤200K | >200K | |
|---|---|---|
| Input | ~$2 | ~$4 |
| Output | ~$12 | ~$18 |
When to Use Pro
- Hardest reasoning — novel logic, graduate-level science, complex multi-constraint problems.
- Huge context — up to 2M tokens for whole-repo or long-document analysis (mind the 200K price cliff).
- Autonomous agents — strong agentic reliability for long tool-using runs.
- Multimodal input — text, image, audio, video in one model.
When Flash Beats Pro
Here's the money-saving truth: for most production traffic, you don't need Pro. Gemini 3.5 Flash (~$1.50/$9) launched beating 3.1 Pro on several coding/agentic benchmarks, and Flash-Lite ($0.10/$0.40) handles classification and routing for pennies. Reserve 3.1 Pro for the genuinely hard reasoning or the 2M-context jobs that Flash can't do. The cascade — Flash-Lite/Flash by default, Pro on escalation — is the biggest Gemini cost lever. See how to reduce AI API costs.
How to Call It
curl -X POST https://api.linkmodel.ai/api/v1/... \
-H "Authorization: Bearer $LINKMODEL_API_KEY" -H "Content-Type: application/json" \
-d '{ "model": "gemini-3.1-pro-preview", "messages": [{"role":"user","content":"Analyze this 300-page contract for risk clauses: ..."}] }'Confirm the exact chat endpoint and schema in the docs. Watch context size — RAG pipelines that push past 200K silently double per-token cost.
Pro vs the Field
At ~$2/$12, Gemini 3.1 Pro undercuts Claude Opus 4.8 ($5/$25) and GPT-5.5 ($5/$30) while offering the largest context. It's the value option at the frontier — full comparison in Claude vs GPT vs Gemini vs DeepSeek and best LLM API.
Bottom Line
Gemini 3.1 Pro is the value frontier reasoner with the biggest context — but try Gemini 3.5 Flash first; it often matches Pro on coding at far less cost.
Start free with a $1 credit and compare Pro vs Flash on your prompts.