Gemini 3 Flash Preview

Gemini 3 Flash is Google DeepMind's Gemini 3 series Flash-tier model, rolled out around mid-December 2025, with the preview being the early developer-access release before general availability. It pairs Pro-grade reasoning with Flash-level speed at less than a quarter the cost of Gemini 3 Pro, and is natively multimodal with a 1M-token input context window and up to 64K output tokens. It posts frontier-level scores on PhD-level reasoning benchmarks such as GPQA Diamond (about 90.4%) while serving coding, document analysis and agentic use cases. Ideal for high-throughput tasks that still need strong reasoning.

curl https://api.tunerouter.com/v1/chat/completions \
  -H "Authorization: Bearer $TUNEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3-flash-preview", "messages": [{"role": "user", "content": "Hello!"}]}'

Model ID: gemini-3-flash-preview · works with OpenAI SDKs — just point base_url at api.tunerouter.com