> ## Documentation Index
> Fetch the complete documentation index at: https://docs.apiyi.com/llms.txt
> Use this file to discover all available pages before exploring further.

# New Gemini_Reverse Group: Gemini Reverse Sources at 0.5x

> APIYI has added a Gemini_Reverse token group backed by Gemini CLI and Antigravity reverse-engineered channels, billed at a 0.5x discount. The default group remains the AI Studio and Vertex official relay. The reverse group is cheaper and suits personal learning and coding; production workloads and stability-sensitive tasks should use the Default official-relay group.

**2026/9/4 23:58 (UTC+8)** · New Model · Google

🚀 **New `Gemini_Reverse` group: a Gemini reverse-engineered economy channel at a 0.5x discount**

Background: `Gemini_Reverse` draws on reverse-engineered channels from Gemini CLI and Antigravity (Reverse Sources) and bills at 0.5x, half the official price. Take `gemini-3.8-flash` as an example: the official price is \$0.75 input / \$3.75 output per 1M tokens, and this group bills at \$0.375 / \$1.875. Select the `Gemini_Reverse` group when creating a token in the console; model names and endpoints are the same as on the default group.

Gemini models now come in two groups, pick by need:

* `default` official-relay group: official AI Studio and Vertex resources, priced line for line with the official rate; for production workloads and tasks that need stability
* `Gemini_Reverse` group: Gemini CLI / Antigravity reverse-engineered resources at a 0.5x discount; for personal learning, coding and other cost-sensitive use, with availability following the reverse sources

See [What are groups? User groups vs token groups](/en/faq/groups-explained) for how groups work, and the [Gemini 3.8 Flash](/en/models/gemini-3-8-flash) page for pricing.

***

← [Back to Live Updates](/en/live) · 📚 [Monthly Archive](/en/live/archive)