Woofun AI reports that B.AI API officially launched Gemini 3.6 Flash and Gemini 3.5 Flash-Lite on July 22. Gemini 3.6 Flash maintains current pricing while reducing average token consumption by 17%, with reductions reaching 65% in complex DeepSWE scenarios. Gemini 3.5 Flash-Lite delivers output speeds up to 350 tokens/s, targeting high-concurrency tasks like document batch processing. Official API interfaces are now open for developer access.