google

Gemini 3.1 Pro Preview

gemini-pro

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

ReasoningTool CallingAttachmentsStructured Output

Context Window

1.0M

Max Output

66K

Temperature

Yes

Open Weights

No

Knowledge Cutoff

2025-01

Released

2026-02-19

Last Updated

2026-02-19

Modalities

Input:TextImageVideoAudioPDF
→
Output:Text

Benchmarks

BenchmarkScoreMetricHarness
SWE-Bench Pro54.2resolve rate-
Terminal-Bench70.3success rateTerminus-2
SWE-Bench Pro46.1resolve rate-
SWE-Atlas Codebase QnA13.5scoreMini-SWE-Agent
SWE-Atlas Refactoring33.81scoreGemini CLI
SWE-Atlas Test Writing29.84scoreMini-SWE-Agent
Artificial Analysis Coding Agent Index43average pass@1Gemini CLI
SWE-Atlas Codebase QnA45.6pass@1Gemini CLI
SWE-Bench Pro15.1pass@1Gemini CLI
Terminal-Bench68.3pass@1Gemini CLI

Available Providers (9)

ProviderInput /1MOutput /1MCache Read /1MCache Write /1MReasoningStatus
Vercel AI GatewayVercel AI GatewayRp 35.381Rp 212.281Rp 3.539—effort—
ZenMuxZenMuxRp 35.381Rp 212.281Rp 3.539Rp 79.606effort—
Merge GatewayMerge GatewayRp 35.381Rp 212.281Rp 3.539—effort—
OrcaRouterOrcaRouterRp 70.761Rp 318.421Rp 3.539—effort—
Perplexity AgentPerplexity AgentRp 35.381Rp 212.281Rp 3.539—effort—
Kilo GatewayKilo GatewayRp 35.381Rp 212.281——effort—
OpenRouterOpenRouterRp 35.381Rp 212.281Rp 3.539Rp 6.634effort—
FastRouterFastRouterRp 35.381Rp 212.281——effort—
NanoGPTNanoGPTRp 35.381Rp 212.281Rp 3.539—effort—
← Kembali ke Katalog