Last updated: August 22, 2026 — verified against Google’s current Gemini API model and pricing documentation.
Google’s Gemini lineup has changed quickly in 2026. The most important update is that Gemini 3.7 Flash, released as generally available on August 13, is now Google’s newest and most capable Flash model for coding and agentic workflows. For high-volume low-cost work, Gemini 3.5 Flash-Lite remains the cost-efficient option, while Gemini 3.1 Pro Preview is still available for Pro-tier reasoning workloads. This guide compares the current choices using Google’s own documentation.
Current Gemini Models in August 2026
- Gemini 3.7 Flash: Google’s newest GA Flash model, built for complex coding, agentic workflows and reliable multi-step execution.
- Gemini 3.6 Flash: the previous-generation Flash model, still listed as stable.
- Gemini 3.5 Flash: a legacy Flash model for routine high-throughput workloads.
- Gemini 3.5 Flash-Lite: Google’s fastest and most cost-efficient 3.5 model for high-volume execution.
- Gemini 3.1 Pro Preview: a Pro-tier model available through the Gemini API for complex reasoning and multimodal tasks.
Google’s own model catalog now recommends Gemini 3.7 Flash as the latest Flash choice. That means older comparisons centered on Gemini 3.1 Pro versus 3.5 or 3.6 Flash need to be read as snapshots rather than a description of the current top Flash model.
Gemini 3.7 Flash: The Current Workhorse
Google describes Gemini 3.7 Flash as its most capable Flash model for agentic workflows and multimodal reasoning. It supports a 1,048,576-token input context, up to 65,536 output tokens, configurable low/medium/high thinking levels, function calling, code execution, search grounding, file search, structured outputs and preview computer use.
For most new coding and agent workloads, 3.7 Flash is the logical first model to benchmark. Google’s migration guidance specifically points users of 3.6 Flash, 3.5 Flash, Gemini 3 Flash Preview and 3.1 Pro toward 3.7 Flash where appropriate.
Gemini API Pricing
| Model | Input / 1M tokens | Output / 1M tokens |
|---|---|---|
| Gemini 3.7 Flash — introductory through Dec. 31, 2026 | $0.75 | $3.75 |
| Gemini 3.7 Flash — from Jan. 1, 2027 | $1.50 | $7.50 |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 |
| Gemini 3.1 Pro Preview — under 200k-token prompts | $2.00 | $12.00 |
Prices above reflect Google’s published standard paid-tier rates at the time of this update. Batch, priority, caching and grounding charges can differ, so production budgets should be checked against Google’s live pricing page.
Which Gemini Model Should You Use?
- Complex coding and agentic workflows: start with Gemini 3.7 Flash.
- High-volume translation, classification and simple processing: Gemini 3.5 Flash-Lite is designed for cost efficiency.
- Workloads specifically requiring Pro behavior: benchmark Gemini 3.1 Pro Preview against 3.7 Flash rather than assuming Pro will automatically be better.
- Existing 3.5/3.6 Flash deployments: review Google’s migration guidance before upgrading because some sampling parameters differ in newer models.
Is Gemini Ultra a Model?
Do not confuse Google’s consumer plan names with Gemini API model IDs. The API model catalog uses names such as Gemini 3.7 Flash, 3.5 Flash-Lite and 3.1 Pro Preview. When comparing models for development, use the actual model IDs in Google’s API documentation rather than subscription branding.
Gemini vs Claude and GPT-5.6
Claude, Gemini and OpenAI now all offer multiple cost-performance tiers. The fairest comparison is task-specific: benchmark the exact coding, reasoning, latency and multimodal workload you care about using current versions and current prices. Model families are moving too quickly for a permanent universal ranking.
FAQ
What is Google’s newest Gemini Flash model? Gemini 3.7 Flash. Google made it generally available on August 13, 2026 and describes it as its most capable Flash model for coding and agentic workflows.
Is Gemini 3.7 Flash production-ready? Yes. Google’s latest-model documentation marks it GA and ready for production use.
What is the cheapest current Gemini model for high-volume work? Google describes Gemini 3.5 Flash-Lite as its most cost-efficient GA model, with standard paid pricing of $0.30 per million input tokens and $2.50 per million output tokens.
