Google DeepMind released three new Gemini AI models on Tuesday — 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber — but the widely anticipated update to its flagship Gemini Pro model missed the June deadline, wiping an estimated $200 billion from Alphabet's market cap. The stock fell 4.4% after Bloomberg reported internal struggles with coding benchmarks.
The new models emphasize efficiency and cost reduction. Gemini 3.6 Flash reduces output token usage by 17% compared to 3.5 Flash, with a price cut from $9 to $7.50 per million output tokens. It scored 49% on the DeepSWE v1.1 software engineering benchmark, up from 37% for 3.5 Flash, and led the OSWorld-Verified test with 83.0%. However, rivals like GPT-5.6 Luna and Claude Sonnet 5 still outperform it on certain coding and knowledge work tasks.
Gemini 3.5 Flash-Lite is built for high-volume, cost-sensitive workflows, delivering 350 output tokens per second at $0.30 per million input tokens. The restricted Gemini 3.5 Flash Cyber will only be available to governments and vetted partners for vulnerability detection.
The missing Gemini 3.5 Pro — teased at Google I/O 2026 for a one-month rollout — is now in partner testing with no firm release date. Product lead Logan Kilpatrick said the team has begun pre-training for Gemini 4, described as its “most ambitious” yet.
For the crypto market, the delay underscores the intense AI competition, where decentralized AI projects could see increased attention as centralized tech giants face hurdles. No specific cryptocurrency is directly affected, but the broader AI narrative may influence sentiment toward AI-themed tokens in the long term.
