Official product status
Google has named Gemini 3.5 Pro and says it is currently testing with partners, with broader availability planned when the model is ready.
Google says Gemini 3.5 Pro is testing with partners and will become broadly available when ready. It is not listed in the public Gemini API catalog, so there is no confirmed public model ID, price, input-mode specification, or callable EvoLink route yet.
EvoLink will notify you after a public upstream route, request model ID, pricing, and route behavior are verified. This is an independent product alert, not Google Early Access, a partner program, or a guarantee of access.
EvoLink product alerts only. No Google affiliation, partner access, access guarantee, or spam.
Want faster updates? Join the EvoLink Discord
The confirmed request model ID, supported input modes and tools, live EvoLink pricing, limits, compatibility, and a controlled migration path through the unified API.
These are the facts needed for a real integration. Google has confirmed partner testing, but the public API details below remain unpublished as of August 1, 2026.
Google has named Gemini 3.5 Pro and says it is currently testing with partners, with broader availability planned when the model is ready.
No public Gemini API, Google AI Studio, Vertex AI, or EvoLink route is documented for Gemini 3.5 Pro. Partner testing is not public Early Access.
Google's public model catalog does not list a Gemini 3.5 Pro request ID. Treat strings such as `gemini-3.5-pro` as unverified placeholders, not production configuration.
No official input, output, caching, batch, priority, or long-context price has been published for Gemini 3.5 Pro. EvoLink will reuse its live pricing module after SKUs are approved.
Text, image, audio, video, PDF, tool calling, context length, output limit, quotas, and regions are not yet documented for the public route.
Prepare the evaluation and routing layer now without making an unreleased model a production dependency.
Store the selected model in configuration so the verified request ID can be added without changing application business logic.
Build a fixed set of coding, agent, document, and multimodal tasks using inputs you are permitted to test.
Measure task success, latency, token use, retries, and tool errors on Gemini 3.1 Pro or Gemini 3.5 Flash now.
Choose the minimum improvement in quality, reliability, or cost per successful task required before moving traffic.
Evaluate the verified route offline, then with shadow requests and a small canary before changing the default model.
Retain a callable Gemini route and explicit timeout, retry, and rollback rules until the new route proves stable.
The reason is operational flexibility, not early access. EvoLink can add the verified Gemini 3.5 Pro route to the same gateway teams already use for available Gemini and cross-provider models.
Keep authentication and request handling stable while comparing Gemini 3.5 Pro with available Gemini, OpenAI, Anthropic, and other routes.
Use live route pricing together with token use, retries, latency, and acceptance results instead of choosing from list price alone.
Move a small share of traffic after validation, keep the previous route available, and roll back without rebuilding a provider-specific integration.
Follow the status evidence, compare the available Flash route, or start with the current Pro option.
Official signals, partner-testing status, unconfirmed claims, and public API checkpoints.
Compare the unavailable Pro route with the callable Gemini 3.5 Flash route without inventing Pro specifications.
Use the current Pro-family route on EvoLink while keeping future model selection configurable.
Decide whether to deploy GPT-5.6 now, wait for Gemini, or prepare a low-risk switch through one gateway.
Yes. Google says Gemini 3.5 Pro is currently testing with partners and will become broadly available when ready.
No. As of August 1, 2026, it is not listed in Google's public Gemini API model catalog, pricing page, or API reference.
Not yet. EvoLink will list the route only after upstream access, the request model ID, pricing, route behavior, and a successful real call are verified.
No public request model ID is confirmed. Do not infer an ID from this page's URL or hard-code a rumored string in production.
Pricing is not published. This page will show EvoLink's live pricing module after the route and SKUs are approved rather than guessing a price.
Google has not published the public route's input modalities, context limit, tool support, output limit, quotas, or regions.
No. Partner testing is restricted testing. Google has not published a public signup, waitlist, eligibility policy, preview route, or access guarantee.
Use Gemini 3.1 Pro for the current Pro-family option or Gemini 3.5 Flash for the publicly available 3.5-generation route on EvoLink.
Keep the model ID configurable, benchmark a fixed task set, add the verified route in shadow mode, move a small canary, and retain a tested fallback before broader rollout.