Skip to content
News Model release

Gemini 3.5 Pro misses its June 30 GA deadline, now targeting July with no firm date

· by Pondero Newsdesk

The short version

Google's Gemini 3.5 Pro failed to reach public general availability by June 30, 2026, the date Google committed to at its I/O developer conference. The model remains in a limited Vertex AI enterprise preview. A Polymarket contract tracking the outcome settled at 97% No-Release, on $229,697 in trading volume.

Gemini 3.5 Pro misses its June 30 GA deadline, now targeting July with no firm date

Gemini 3.5 Pro did not ship to the public by June 30, 2026, the specific date Google had tied to the model at its I/O conference in May. A Polymarket prediction market for the outcome closed at 97% probability of No-Release on $229,697 in trading volume. The model remains in a limited Vertex AI enterprise preview as of today.

What happened

Google confirmed the delay publicly. Per the company's stated rationale, engineers determined that early tester feedback showed excessive token consumption during extended agentic tasks, and that additional validation work on long-horizon performance was needed before a public rollout. Google has set July as the new target but has not committed to a specific date within that month per Bind AI's coverage.

The expected specifications for Gemini 3.5 Pro are a 2-million-token context window, a Deep Think reasoning mode restricted to Ultra subscribers at $250 per month, frontier-class multimodal capability, and pricing estimated at approximately $15 per million input tokens and $60 per million output tokens, making it roughly 10x the cost of Gemini 3.5 Flash.

The June miss is the second consecutive Google I/O commitment to slip. Gemini Ultra 1.5 experienced a comparable delay earlier in 2026.

Why it matters

The real story is the talent signal: four senior Gemini researchers announced departures to Anthropic and OpenAI the same week Google postponed the launch. Four senior Gemini researchers announced departures to Anthropic and OpenAI during the week of June 21-27, the same week Google postponed the public launch per Bind AI. The combined signal, a missed public commitment arriving alongside a talent outflow, wiped roughly $270 billion from Alphabet's market cap at the time.

For operators who had workloads queued for Gemini 3.5 Pro's 2-million-token window, the delay is a concrete planning problem. Claude Opus 4.8 currently claims a 1-million-token context. Gemini 3.5 Pro's window, per Google's own preview specs, would double that. Every additional month in preview is a month that window is only available to the handful of organizations inside the Vertex AI early-access program, not to the developer market broadly.

The pattern also intersects with the US government's voluntary frontier AI review framework. Gemini 3.5 Pro is expected to be among the first Google models subject to that process per the June 2 White House executive order. If the review is part of what is extending the timeline before public launch, that is a material data point about how the framework will affect future release cadences from all frontier labs, not just Google.

Context

June 2026 closed with no major AI lab delivering a publicly accessible frontier model on its originally stated schedule. OpenAI previewed GPT-5.6 Sol, Terra, and Luna on June 26 but restricted access to roughly 20 government-vetted organizations, not a public launch. xAI's Grok 5 remained in training on its Colossus 2 cluster, with Polymarket contracts closing near 3% probability of a June release. The simultaneous slippage across four anticipated releases points to external constraints, government review timelines and regulatory caution, as much as engineering schedules.

Google's position is structurally more exposed than the others. The company has historically used I/O to anchor concrete launch commitments, and two consecutive misses against I/O promises starts to look like a credibility gap on the roadmap. Anthropic and OpenAI have been more conservative about public deadline statements at developer conferences, which gives them less to miss against.

The talent exodus compounds the technical narrative. Losing four senior Gemini researchers in a single week, going to the two companies that are most visibly winning the enterprise adoption race, is the kind of signal that enterprise procurement teams will notice when they model multi-year AI platform risk.

What to watch next

Three things carry the most decision-relevance in July. First, whether Google names a specific date for Gemini 3.5 Pro GA rather than maintaining the open "July target." A date would signal the validation work has concluded. Second, whether the company acknowledges the White House review process as a factor in the delay or keeps to the "tester feedback" framing. Third, whether the research team departures continue. A second wave would accelerate the story from a deployment story into a capability-trajectory story.

For developers who need the 2-million-token window before the end of Q3, the practical path today is applying for Vertex AI enterprise preview access. That program is running with active users; a July GA date would widen access substantially.

Sources