Google DeepMind shipped three Gemini models on Tuesday, 3.6 Flash, 3.5 Flash-Lite, and a cybersecurity-tuned 3.5 Flash Cyber available only to governments and trusted partners through a limited-access pilot on CodeMender. What it didn’t ship, again, was Gemini 3.5 Pro. The flagship’s last update was February, and back in May Google told users it was “already being used internally, and we look forward to rolling it out next month.” That was two months ago.
The Flash releases are respectable on their own terms. Product lead Tulsee Doshi’s post frames 3.6 Flash as a real efficiency step: 17% fewer output tokens than 3.5 Flash, a jump from 37% to 49% on the DeepSWE coding benchmark, and a knowledge cutoff advanced from January 2025 to March 2026. Pricing lands at $1.50 per million input tokens and $7.50 per million output. Flash-Lite runs at 350 output tokens per second for $0.30 in and $2.50 out.
But the market isn’t grading Flash. Bloomberg’s Julia Love and Davey Alba reported on July 16 that Alphabet is months behind on Gemini 3.5 Pro, with coding performance falling short of internal targets and a late-June attempt to fix things by refreshing training data producing disappointing results. Alphabet stock fell roughly 4.4% on the report, erasing about $200 billion in market cap in a session.
That’s the context Tuesday’s launch is actually operating in. TechCrunch’s Rebecca Bellan notes OpenAI has already released GPT-5.5 and is rolling out GPT-5.6, Anthropic has shipped Claude Opus 4.8, Claude Sonnet 5, and expanded access to Fable 5. Gizmodo observes that 3.6 Flash lags Claude Sonnet 5 and GPT-5.6 across most benchmarks and even trails Grok 4.5 in agentic coding.
DeepMind product lead Logan Kilpatrick said 3.5 Pro is in partner testing and will “land soon,” and that the team has begun its most ambitious pre-training run yet, for Gemini 4. The subtext is familiar from every quarterly-narrative moment in Big Tech going back to the 2022 Meta pivot: when the current generation isn’t landing, point at the next one.
Sources
- https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/
- https://techcrunch.com/2026/07/21/google-releases-three-new-gemini-models-but-no-3-5-pro/
- https://9to5google.com/2026/07/21/gemini-3-6-flash-launch/
- https://www.bloomberg.com/news/articles/2026-07-16/google-gemini-launch-delayed-as-tech-falls-short-of-internal-goals
- https://gizmodo.com/google-introduces-gemini-3-6-to-remind-you-it-has-an-ai-model-too-2000788641