Googlemultimodal

Gemini 3.1

Name: Gemini 3.1
Brand: Google
Price: 2 USD
Availability: InStock

Gemini 3.1 is Google DeepMind's latest frontier model, pushing the boundaries on multimodal reasoning. It builds on the Gemini 2.5 Pro thinking architecture with significant improvements in factuality and agentic capabilities.

Released

Mar 2025

Context

Input Price

$2/1M

Output Price

$8/1M

Specifications

Parameters: Undisclosed

Context Window: 2,097,152 tokens

Open Source: No

Category: multimodal

Benchmarks

MMLU93%

HumanEval94.5%

MATH93.2%

GPQA82%

Capabilities

Advanced reasoningMultimodal understandingAgentic workflowsCode generationResearchTool use

✓Vision

✓Function Calling

✓Streaming

✗Fine-tuning

Pros

✓Frontier-level performance
✓2M context window
✓Strong agentic capabilities
✓Competitive pricing

Cons

✗Very new — limited production track record
✗Google ecosystem preferred
✗No fine-tuning yet
✗Rate limits at launch

How It Compares

See how Gemini 3.1 stacks up against the competition. View all comparisons →

Best For

Google's frontier model. Competitive with GPT-5 and Claude 4 Opus across the board.