AI providers this week: Gemini 2.5 deprecation

Transcript

Gemini 2.5 Pro and Flash requests will reroute to newer models on December 1, 2026. Plus Embedding 2 preview, GPT-Image-2.6 models, Message threads, and four more AI provider updates this week.

Gemini 2.5 Pro and Flash requests will be billed at the rates of Gemini 3.8 Flash and Gemini 3.5 Flash-Lite, so plan for that.

The Gemini embedding 2 preview model is deprecated, and since November 6, 2026 its requests route to Gemini embedding 2 at its rates.

GPT-Image-2.6 Sunburst and Flare are now generally available, and they can generate SVG and animated GIFs as well as edit images.

Message threads, now in preview, keep your conversation on the server so later requests send only new messages.

The Claude Console IP allowlist, now in preview, refuses requests from unlisted addresses, with up to 1,000 entries per organization.

GPT-6-Cyber is a model for investigating vulnerabilities, reproducing them in authorized test environments and verifying fixes.

Requests to grok-imagine-image-quality and grok-imagine-image-pro are now served by grok-imagine-image-2.0-low at a lower per-image price.

Live API outbound SIP calling is now open to all organizations, connecting GPT-Live to a phone number.

That's your week in AI providers. The one to act on: check your Gemini 2.5 usage before December 1, 2026, when billing moves to the new models' rates.

Follow for next week's AI provider updates. The full list is at cloud changelog dot com.