--- title: "gemini-2.5-pro on Gemini API retirement date | Since.dev" description: "gemini-2.5-pro on Gemini API is routed to gemini-3.8-flash on December 1, 2026. Replacement: gemini-3.8-flash." canonical_url: https://since.dev/calendar/google/gemini-2.5-pro/gemini last_updated: 2026-11-12 --- # gemini-2.5-pro on Gemini API is routed to gemini-3.8-flash on December 1, 2026 AI model - **Date:** December 1, 2026 (UTC) - **Replacement:** `gemini-3.8-flash` - **Platform:** Gemini API - **Provider:** [Google](https://since.dev/calendar/google.md) - **Last checked:** Nov 14, 2026 gemini-2.5-pro Routes to gemini-3.8-flash December 1, 2026 on Gemini API In its place: `gemini-3.8-flash` since.dev/calendar/google/gemini-2.5-pro/gemini ## What changes Requests keep working, answered by gemini-3.8-flash. ## What works differently on `gemini-3.8-flash` From its documentation. Moving to it is more than a new model ID. - Default thinking level is MEDIUM; supported thinking levels are LOW, MEDIUM, HIGH. To keep the old behavior: set thinking_level explicitly if the code relied on a different reasoning level. [Source](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-3-8-flash) - Explicitly setting thinking_level to MINIMAL returns an API validation error. To keep the old behavior: do not set thinking_level to MINIMAL; use LOW, MEDIUM, or HIGH. [Source](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-3-8-flash) - Prefilled model turns are unsupported and must be removed when upgrading. To keep the old behavior: remove prefilled model turns from conversation histories. [Source](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-3-8-flash) - Integer thinking_budget configurations are replaced by the string enum thinking_level. To keep the old behavior: replace thinking_budget with the string enum thinking_level. [Source](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-3-8-flash) - Passing frequency_penalty or presence_penalty throws an active API error. To keep the old behavior: remove frequency_penalty and presence_penalty from legacy client code. [Source](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-3-8-flash) - History payloads cannot end with a model role turn. To keep the old behavior: ensure conversation history does not end with a model role turn. [Source](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-3-8-flash) - A FunctionResponse turn must strictly match the id, name, and execution count of the preceding FunctionCall. To keep the old behavior: make each FunctionResponse match the id, name, and execution count of the preceding FunctionCall. [Source](https://docs.cloud.google.com/gemini-enterprise-agent-platform/models/guides/gemini-3-8-flash) ## Source [aistudio.google.com](https://aistudio.google.com/docs/deprecations) > gemini-2.5-pro | June 17, 2025 | December 1, 2026 | gemini-3.8-flash (gemini-2.5-pro is no longer available in Google AI Studio; effective December 1, 2026, API requests to gemini-2.5-pro will automatically route to gemini-3.8-flash) Data license: [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). From Google's documentation. Quotes are shortened and dates restated. ## Does your code use it? Paste a public repository. Since.dev analyzes it and shows the files and lines these dates hit. [Add to your calendar](webcal://api.since.dev/v1/calendar.ics?provider=google) · [iCal file](https://api.since.dev/v1/calendar.ics?provider=google) · [RSS](https://api.since.dev/v1/calendar.rss?provider=google) ## More from Google - Nov 19, 2026 · no sooner than · [gemini-3.6-flash on Vertex AI](https://since.dev/calendar/google/gemini-3.6-flash/vertex.md) · Google · AI model · replace with gemini-3.8-flash - Dec 1, 2026 · [gemini-2.5-flash on Gemini API](https://since.dev/calendar/google/gemini-2.5-flash/gemini.md) · Google · AI model · replace with gemini-3.5-flash-lite - Dec 3, 2026 · [gemini-omni-1.1-flash on Gemini API](https://since.dev/calendar/google/gemini-omni-1.1-flash/gemini.md) · Google · AI model · replace with gemini-omni-1.2-flash - Dec 13, 2026 · no sooner than · [gemini-live-2.5-flash-native-audio on Vertex AI](https://since.dev/calendar/google/gemini-live-2.5-flash-native-audio/vertex.md) · Google · AI model · replace with gemini-3.8-live - Dec 18, 2026 · no sooner than · [gemini-2.5-pro-preview-tts on Gemini API](https://since.dev/calendar/google/gemini-2.5-pro-preview-tts/gemini.md) · Google · AI model · replace with gemini-3.8-flash-tts or gemini-3.8-flash-lite-tts - Dec 18, 2026 · no sooner than · [gemini-2.5-flash-preview-tts on Gemini API](https://since.dev/calendar/google/gemini-2.5-flash-preview-tts/gemini.md) · Google · AI model · replace with gemini-3.8-flash-tts or gemini-3.8-flash-lite-tts ## Put your code on continuous maintenance. Scan a public repository now, or give your agent one line. [Scan a repository](https://since.dev/scan.md). Any public GitHub repository. No sign-in. ### Tell your agent what changed since it was trained. `claude mcp add --transport http since https://api.since.dev/mcp` [Get an API key](https://app.since.dev/login?mode=signup) Codex, Cursor and any MCP client use https://api.since.dev/mcp. --- Every page on this site as Markdown: https://since.dev/sitemap.md. For agents: https://since.dev/llms.txt.