Gemini 3.5 Pro remains in testing after reported delay

Bloomberg reported on July 16 that Gemini 3.5 Pro was months behind schedule as Google worked to improve its capabilities, particularly coding. 9to5Google's coverage of the report noted that Google had previously pointed to a June arrival and that the company was still testing the model with partners.

Those details come from reporting about Google's internal development process. They do not establish that the model was cancelled, and Google has not published a replacement launch date. Claims about internal training changes or employee concerns should therefore remain attributed to the Bloomberg report rather than presented as confirmed product documentation.

Google gave a first-party status update on July 21 while announcing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. The company said Gemini 3.5 Pro was currently testing with partners and that it planned to make the model broadly available as soon as it was ready. The post did not provide a date, API availability, pricing, or a model card for 3.5 Pro.

Axios separately noted that the July 21 releases did not include 3.5 Pro and connected the omission to Bloomberg's delay report. Together, the sources support a narrow conclusion: a delay was reported, and Google's latest public statement confirms continued testing without committing to a schedule.

Engineering teams should treat any migration plan as provisional until Google publishes the API surface, availability tiers, pricing, context limits, and safety documentation. Existing production workloads do not need to move merely because a future flagship model has been discussed.

When 3.5 Pro becomes available, compare it on repository-specific coding tasks, tool-use reliability, long-horizon agent workflows, latency under concurrency, and total token cost. Keeping prompts, evaluations, and orchestration portable will make that comparison easier without assuming that an unreleased model will meet a particular quality or cost target.