NOUMENOpen app
NOUMEN

Google · Frontier

Gemini 3.1 Pro (Preview)

Added Feb 19, 2026

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Gemini 3.1 Pro preview is built for tasks where simple answers are not enough. Stronger core reasoning for complex coding, math, and long-context workflows, with multimodal support and a reported 77.1% verified score on ARC-AGI-2. NOTE: Inputs > 200k tokens are charged at 2x input and 1.5x output rates.

  • Reasoning
  • Tools
  • Vision
  • Structured output
  • Audio input
  • Video input
Input
Text, Images, Video, Audio
Output
Text
Reasoning effort
low · medium · high
Discovery tier
Frontier

Pricing

USD per 1M tokens
Input
$2.00
Output
$12.00
Cache read
$0.20
Cache write
$0.38

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
65.5K
Average output (7 days)
699 tokens
Longer than 62% of 465 models

Side by side

Gemini 3.1 Pro (Preview) compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Gemini 3.1 Pro (Preview)GoogleThis modelGemini 2.5 Pro Preview 0325GoogleRemove
ProviderGoogleGoogle
Discovery tierFrontierBalanced
AddedFeb 19, 2026Mar 25, 2025
Limits
Context window1.05M1.05M
Max output65.54K65.54K
Average output699 tokens1.3K tokens
Pricing, USD per 1M tokens
Input$2.00Best$2.50
Output$12.00$10.00Best
Cache read$0.20Best$0.25
Cache write$0.38—
Performance
Throughput68.0 tok/s—
Time to first token4.7 s—
Request success (24h)99.8%—
Benchmark indexes
Intelligence Index29.7Best16.1
Coding Index68.8Best33.3
Math Index—87.7
Capabilities
ReasoningYesYes
ToolsYesNo
VisionYesYes
Structured outputYesNo
Audio inputYesYes
Video inputYesNo
Reasoning effortlow, medium, highlow, medium, high

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Gemini 3.1 Pro (Preview)
  • Gemini 2.5 Pro Preview 0325

Intelligence Index

  • Gemini 3.1 Pro (Preview)29.7
  • Gemini 2.5 Pro Preview 032516.1

Gemini 3.1 Pro (Preview): Better than 71% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Gemini 3.1 Pro (Preview): 68.8Top
    • Gemini 2.5 Pro Preview 0325: 33.3

  • Math Index

    Composite of the math evaluations

    • Gemini 3.1 Pro (Preview): —
    • Gemini 2.5 Pro Preview 0325: 87.7

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Gemini 3.1 Pro (Preview): 35.4%Top
    • Gemini 2.5 Pro Preview 0325: 2.2%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Gemini 3.1 Pro (Preview): 9.4%Top
    • Gemini 2.5 Pro Preview 0325: 0.0%

  • Harvey LAB-AAv1.1

    Legal tasks passed in full, with no material hallucination

    • Gemini 3.1 Pro (Preview): 0.0%
    • Gemini 2.5 Pro Preview 0325: —

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Gemini 3.1 Pro (Preview): 456 EloTop
    • Gemini 2.5 Pro Preview 0325: 302 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Gemini 3.1 Pro (Preview): 791 EloTop
    • Gemini 2.5 Pro Preview 0325: 455 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Gemini 3.1 Pro (Preview): 17.8%Top
    • Gemini 2.5 Pro Preview 0325: 10.2%

  • AA-LCRv1.1

    Reasoning across long documents

    • Gemini 3.1 Pro (Preview): 82.0%Top
    • Gemini 2.5 Pro Preview 0325: 69.0%

  • MLCR-AA

    Long-context reasoning over medical records

    • Gemini 3.1 Pro (Preview): 11.1%
    • Gemini 2.5 Pro Preview 0325: 12.2%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Gemini 3.1 Pro (Preview): 47.0%Top
    • Gemini 2.5 Pro Preview 0325: 22.5%

  • GPQA Diamond

    Graduate-level science questions

    • Gemini 3.1 Pro (Preview): 94.1%Top
    • Gemini 2.5 Pro Preview 0325: 84.4%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • Gemini 3.1 Pro (Preview): —
    • Gemini 2.5 Pro Preview 0325: 86.2%

  • IFBench

    Following precise output instructions

    • Gemini 3.1 Pro (Preview): 77.1%Top
    • Gemini 2.5 Pro Preview 0325: 48.7%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Gemini 3.1 Pro (Preview): 4.0%Top
    • Gemini 2.5 Pro Preview 0325: 0.0%

  • SciCode

    Scientific computing in Python

    • Gemini 3.1 Pro (Preview): 58.7%Top
    • Gemini 2.5 Pro Preview 0325: 46.3%

  • LiveCodeBench

    Recent competitive programming problems

    • Gemini 3.1 Pro (Preview): —
    • Gemini 2.5 Pro Preview 0325: 80.1%

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • Gemini 3.1 Pro (Preview): —
    • Gemini 2.5 Pro Preview 0325: 87.7%

  • AIME

    Competition mathematics

    • Gemini 3.1 Pro (Preview): —
    • Gemini 2.5 Pro Preview 0325: 88.7%

  • MATH-500

    Competition math problems

    • Gemini 3.1 Pro (Preview): —
    • Gemini 2.5 Pro Preview 0325: 96.7%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Gemini 3.1 Pro (Preview): 53.8%Top
    • Gemini 2.5 Pro Preview 0325: 26.5%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Gemini 3.1 Pro (Preview): 95.6%Top
    • Gemini 2.5 Pro Preview 0325: 54.1%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Gemini 3.1 Pro (Preview): 82.0%Top
    • Gemini 2.5 Pro Preview 0325: 69.0%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
68.0 tok/s
Faster than 45% of 250 models
Time to first token
4.7 s
Quicker than 36% of 250 models
Prompt cache hits
28.0%
About $1.50 per 1M input tokens with caching
Request success (24h)
99.8%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 97.2%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-3.1-pro-preview",
    "reasoning_effort": "medium",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="google/gemini-3.1-pro-preview",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Google

Other Google models, starting with the same family.