Gemini 3 Pro

GoogleProprietary

Google's frontier model — top-tier math and multimodal, 1M context, aggressive pricing. One of the few models with observed positive drift.

Index @ High

79.0

Context

$/M in · out

$2 · $12

Median speed

130 tok/s

Released

Nov 18, 2025

02·Which effort for which task

The sweet spot, per task

Cheapest effort tier that keeps ≥96% of this model's peak quality on the benchmarks behind each task type. Below the floor number, quality is not okay — it just fails quietly.

Everyday writing & Q&A

Medium

Chat, drafting, summaries, general questions.

keeps 97.5% of peak−87% cost

Code assist

High

Autocomplete, single functions, small fixes.

keeps 98.3% of peak−62% cost

Coding agents

High

Real repos, multi-file changes, end-to-end tickets.

keeps 96.5% of peak−62% cost

⚠ at Low: only 78% of peak — avoid

Math & quantitative

X-High

Derivations, statistics, financial models.

keeps 100.0% of peakpeak tier needed

⚠ at Low: only 66% of peak — avoid

Hard research problems

High

Novel, frontier, genuinely difficult questions.

keeps 96.0% of peak−62% cost

⚠ at Low: only 75% of peak — avoid

Agents & tool use

High

Long workflows with tools, policies, and state.

keeps 97.8% of peak−62% cost

03·The effort ladder

Quality and cost, tier by tier

On everyday tasks, Medium keeps 97.5% of X-High quality at $0.040 vs $0.31 per task.

Intelligence Index by effort

Cost per typical task by effort

04·Benchmark detail

Scores at Low

Low Medium High X-High

at Lowat X-High (peak)

Benchmark	Low	Peak	What it predicts
MMLU-Pro	82.7	88.0	Everyday Q&A · Drafting emails & docs · Summarizing articles
GPQA Diamond	68.3	91.0	Technical analysis · Root-cause debugging · Scientific writing
Humanity's Last Exam	25.5	34.0	Novel research questions · Frontier problem solving
AIME 2025	56.3	97.0	Financial modeling · Statistics · Algorithm design
SWE-bench Verified	61.6	77.0	Coding agents · Bug fixing in real repos · Multi-file refactors
LiveCodeBench	67.2	84.0	Writing functions from scratch · Algorithms · Code autocomplete
τ²-bench	64.7	84.0	Tool-using agents · Customer-facing automation · Multi-step workflows
IFEval	87.4	92.0	Structured output (JSON) · Templated generation · Batch pipelines

06·Drift

Is it still the same model this week?

Weekly Intelligence Index · dashed rules mark serving-change events

Mar 8, 2026Google AI+0.9 ptsstill in effect

Silent checkpoint update — math and coding scores improved without an announcement. Drift cuts both ways.

07·Call it

One key, this model, your effort

The gateway speaks the OpenAI Chat Completions dialect. Pick a host with gemini-3-pro@host syntax, set reasoning_effort, and the sweet spot above becomes one line of config. Full docs →

chat with Gemini 3 Pro

curl https://models.fyi/api/v1/chat/completions \
  -H "Authorization: Bearer $MFYI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3-pro",
    "reasoning_effort": "medium",
    "messages": [{"role": "user", "content": "Hello"}]
  }'
# local dev: replace https://models.fyi with http://localhost:3000