Is Gemini 3.7 Flash available with Zero Data Retention on Elosia?

Gemini 3.7 Flash runs on Elosia's standard backend; ZDR is available on many other models in the catalog.

How much does Gemini 3.7 Flash cost on Elosia?

Gemini 3.7 Flash is billed by usage at $0.375 per 1M input tokens and $1.875 per 1M output tokens, within your Elosia plan.

What is Gemini 3.7 Flash's context window?

Gemini 3.7 Flash supports a context window of up to 1,048,576 tokens.

Is Gemini 3.7 Flash hosted in the EU?

Yes, Gemini 3.7 Flash is available on Elosia's EU-hosted (EdenAI) backend for enterprise organizations.

When should you choose Gemini 3.7 Flash?

Gemini 3.7 Flash is a balanced model: a strong quality/cost trade-off for everyday production work.

Does Gemini 3.7 Flash support image inputs on Elosia?

Yes, Gemini 3.7 Flash accepts image inputs; image tokens are priced separately (see the pricing table).

Gemini 3.7 Flash

GeminiBalanced

ThinkingTool UseVisionStructured Output

About this model

Gemini 3.7 Flash is a multimodal model from Google for fast agentic workflows, coding, and complex multi-step reasoning. It is designed for tasks that require responsive performance and reliable multi-step...

Performance Tier

Balanced

Gemini 3.7 Flash is a balanced model from Gemini : strong performance at a reasonable price.

Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.

Pricing

This model is included in Elosia plans

Affordable

Low cost. Suitable for sustained use and high-volume interactions.

Type	per 1M tokens
Input (prompt)	$0.375
Output (completion)	$1.88
Image	$0.375
Internal reasoning	$1.88
Cache read	$0.037
Cache write	$0.021

Capabilities

Context Length1.0M

Max Output Tokens66K

TokenizerGemini

Inputtext, image, video, file, audio

Outputtext

Release DateAugust 13, 2026

Benchmarks

General Intelligence

MMLU

Not reported

GPQA Diamond

94.5%

Mathematics

MATH-500

Not reported

Programming

HumanEval

Not reported

Reasoning

Humanity's Last Exam

47.9%

Multimodal

MMMU-Pro

85.5%

Agentic

Terminal-Bench 2.1

85.8%

Where does Gemini 3.7 Flash stand?

Compare its performance index against every other model.

View the performance leaderboard

Recommended Use Cases

CodingAnalysisGeneral ChatResearchData Extraction

Strengths

Strongest agentic terminal work in the Flash line (Terminal-Bench 2.1 85.8%, up from 78.0% on Gemini 3.6 Flash)
Large software-engineering gain over its predecessor (DeepSWE v1.1 65.3% vs 48.6%, FrontierCode 1.1 Main 43.6% vs 34.4%)
Leading web development among Flash models (Code Arena 1588 Elo, up from 1538)
Near-perfect long-context retrieval (GDM-MRCR v2 8-needle 97.0% at 128k average); separately, the model supports a 1M-token context window with 64K output
Broad multimodal coverage: video (LVBench 85.4%), computer use (OSWorld-2.0 47.9%)
Best relative standing against rival models on multidisciplinary reasoning (HLE-Verified 53.6%)

Limitations

The only regression versus Gemini 3.6 Flash in Google's table is CharXiv Reasoning (84.5% vs 85.2% without tools, 88.7% vs 89.4% with tools), a step back for chart-heavy analysis and data extraction
Last of the four models Google compares on GDPVal-AA v2 knowledge work (1525 Elo versus 1578 to 1628 for the rival models)
Behind GPT-5.6 Terra on DeepSWE v1.1 (65.3% vs 69.6%), Terminal-Bench 2.1 and 3.0, and OSWorld-2.0 (47.9% vs 50.2%), and behind Claude Sonnet 5 on Agent's Last Exam (26.3% vs 33.3%)
Terminal-Bench 3.0 at 14.9% shows the harder agentic frontier remains largely unsolved
A refinement of Gemini 3.6 Flash rather than a new pretraining run
Google publishes no MMLU, GPQA Diamond or MMMU-Pro figures; the GPQA Diamond, MMMU-Pro and Humanity's Last Exam scores here come from third-party testing (Artificial Analysis)