For Developers/Models/Gemini 3.7 Flash
Closed SourceGoogleReleased 2026-08

Gemini 3.7 Flash

Most intelligent workhorse model yet for coding and agents

Context window

1M

Input / 1M tokens

$0.75

Output / 1M tokens

$3.75

Provider

Google

Data verified 2026-08-14

Gemini 3.7 Flash is a refinement of the 3.6 architecture with algorithmic improvements focused on software engineering, knowledge work, and web development workflows. It delivers substantial gains in coding benchmarks (DeepSWE, FrontierCode) and autonomous agent capabilities at an introductory price of $0.75/1M input tokens through December 31, 2026.

Capability index

Relative estimates (0-100) to place this model against its peers, grounded in published benchmarks.

Coding
8
Reasoning
6
Math
6
Multimodal
8
Long context
9
Speed
9
Cost efficiency
9

How to access it

Available via Gemini API, Google AI Studio, Android Studio, Google Antigravity, Gemini Enterprise Agent Platform, and Gemini Spark (for Google AI Pro and Ultra subscribers in 160+ countries).

Strengths

  • Strong coding performance with significant benchmark improvements
  • Better at multi-step planning and autonomous agent execution
  • Improved debugging and issue resolution capabilities
  • 50% introductory pricing discount through year-end
  • 1M token context window for long-document processing

Best for developers who...

Software engineering and code generationAutonomous agent workflows and business automationDocument-heavy knowledge workWeb developmentHigh-volume coding tasks where cost matters

When to choose it (and when not to)

Reach for Gemini 3.7 Flash when...

  • When building coding agents or multi-step workflows
  • For projects requiring fast execution and lower costs
  • During introductory pricing period (through Dec 31, 2026)
  • For document processing and knowledge work

Look elsewhere if...

  • Complex reasoning tasks requiring Gemini 3.5 Pro
  • After January 1, 2027 (pricing doubles)
  • Computer use / terminal operations (where GPT-5.6 Terra leads)
  • If operating in EEA, UK, Switzerland, or Nigeria (limited Spark availability)

How to use it

  • Leverage its improved instruction-following for multi-step tasks
  • Use for iterative debugging and code refinement workflows
  • Test against 3.6 Flash to verify cost-per-task improvements before migration

Quickstart

Python
import anthropic

client = anthropic.Anthropic(api_key='your-api-key')
message = client.messages.create(
  model='gemini-3.7-flash',
  max_tokens=1024,
  messages=[{'role': 'user', 'content': 'Write a Python function to sort a list'}]
)
print(message.content[0].text)

Note: Use official Google AI SDK instead. This is pseudocode for illustration.

API model id: gemini-3.7-flash

Benchmarks

BenchmarkScoreNotes
DeepSWE v1.165.3%Up from 49.0% for Gemini 3.6 Flash
FrontierCode 1.1 Main43.6%Up from 34.4% for Gemini 3.6 Flash
AutomationBench30.4%Up from 17.0% for Gemini 3.6 Flash
WebDev Arena1588 EloUp from 1538 for Gemini 3.6 Flash
GDP.pdf34.0%Up from 22.0% for Gemini 3.6 Flash

Source: Google Official Announcement

Compare Gemini 3.7 Flash

Compare Gemini 3.7 Flash with any other model

Build a comparison →
All model comparisons →

Learn the concepts