Google DeepMind Nears Gemini 4 Release After Flagship Model Gap
The AMW Read
Gemini 4 is an incremental but consequential update from a major foundation-model player after a prolonged flagship release gap.
Google DeepMind Nears Gemini 4 Release After Flagship Model Gap
Google DeepMind is refining Gemini 4 and intends to release an early post-training version as soon as possible, according to new division leader Koray Kavukcuoglu. In his first media appearance as DeepMind’s leader, Kavukcuoglu said the lab is excited by its results and plans to continue rapid iterations. The target is a launch well before the end of 2026, after Google skipped the Gemini 3.5 Pro update that CEO Sundar Pichai had indicated would arrive in June.
The timing matters because Google has not introduced a new flagship model since the Gemini 3 series in November 2025, while OpenAI’s GPT-6 and Anthropic’s Mythos models have since launched. The reported release plan is therefore less a routine version update than an attempt to restore a competitive flagship cadence after DeepMind shifted attention toward faster, less capable Flash models. The leadership transition adds pressure: Kavukcuoglu is publicly tying his first months in charge to a visible test of whether Google can close a release-cycle gap.
For builders, Gemini 4’s eventual availability could affect model-selection decisions, but the reported details do not establish capabilities, pricing, access terms, or benchmarks. Teams should avoid treating a near-term release signal as a deployment commitment and continue evaluating production models on their own latency, reliability, and cost requirements. For investors, the key evidence will be whether Google follows an initial release with sustained iteration and meaningful developer adoption, rather than merely re-entering the headline cycle. Google DeepMind appeared in 32 pipeline-ingested name-matched items in the last 90 days versus 22 in the prior 90, per the AI Market Watch index.
