Gemini 2.5 Pro Preview 03 25 vs Gemini 3.1 Pro Preview

Wins 0 of 5 benchmarks
$1.30 / task
Wins 0 of 5 benchmarks
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Gemini 2.5 Pro Preview 03 25—
Gemini 3.1 Pro Preview47
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Gemini 2.5 Pro Preview 03 25—
Gemini 3.1 Pro Preview4
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Gemini 2.5 Pro Preview 03 25—
Gemini 3.1 Pro Preview35.4
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Gemini 2.5 Pro Preview 03 25—
Gemini 3.1 Pro Preview1,481
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Gemini 2.5 Pro Preview 03 25—
Gemini 3.1 Pro Preview82.4

Specs

SpecGemini 2.5 Pro Preview 03 25Gemini 3.1 Pro Preview
Input, per 1M!$2.50$2
Output, per 1M!$10$12
Cost per task?—$1.30
Speed, tok/s—115
Context1M1M
WeightsNot verifiedClosed
ReleasedMar 2025Feb 2026
See images✓✓
Read PDFs—✓
Hear audio✓✓
Watch video—✓
Call tools—✓
Force a tool call?Partial
Return strict JSON—✓
Search the web?✓
Run code?✓
Use a computer??

What breaks when you switch

Gemini 2.5 Pro Preview 03 25 → Gemini 3.1 Pro Preview
Cannot use both thinking controls›

Setting thinking_level and legacy thinking_budget together returns HTTP 400; use only one.

Gemini 3.1 Pro Preview → Gemini 2.5 Pro Preview 03 25
No PDF input
No strict JSON