Google

Gemini 3.1 Pro Preview

Reasoning-first Gemini preview for agentic coding and complex problem solving

google/gemini-3.1-pro-preview

Context

1M

Output

65.5K

Released

2026-02-19

Knowledge cutoff

2025-01

Types

InputTextImageAudioVideoPDF
OutputText

Reasoning

Supported

Tool use

Supported

Structured output

Supported

Attachments

Supported

Open weights

—

Provider pricing

Kilo Gateway
InTextImageAudioVideoPDF
OutText
Price
$1.00 / $6.00
Context
1M
Status
Stable
Eden AI
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
FastRouter
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
Impossibl
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
Merge Gateway
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
NanoGPT
InTextImageAudioVideo
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
Ofox
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
OpenRouter
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
OrcaRouter
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
Perplexity Agent
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
Prime Inference
InTextImageAudioVideo
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
Tempr Gateway
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
Vercel AI Gateway
InTextImagePDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
ZenMux
InTextImageAudioVideoPDF
OutText
Price
$2.00 / $12.00
Context
1M
Status
Stable
AI/ML API
InTextImageAudioVideoPDF
OutText
Price
$2.60 / $15.60
Context
1M
Status
Stable

Benchmarks

  • SWE-Bench Pro

    54.2 · resolve rate

  • Terminal-Bench

    70.3 · success rate

  • SWE-Bench Pro

    46.1 · resolve rate

  • SWE-Atlas Codebase QnA

    13.5 · score

  • SWE-Atlas Refactoring

    33.81 · score

  • SWE-Atlas Test Writing

    29.84 · score

  • Artificial Analysis Coding Agent Index

    43 · average pass@1

  • SWE-Atlas Codebase QnA

    45.6 · pass@1

  • SWE-Bench Pro

    15.1 · pass@1

  • Terminal-Bench

    68.3 · pass@1

  • GPQA Diamond

    94.3 · accuracy

  • Humanity's Last Exam

    44.4 · accuracy

  • ARC-AGI-2

    77.1 · accuracy

  • MMMU Pro

    80.5 · accuracy

  • MCP Atlas

    78.2 · success rate

  • OSWorld-Verified

    76.2 · success rate

  • CharXiv Reasoning

    83.3 · accuracy

  • GDPval-AA

    1314 · Elo

From the same lab