skip to content
superbot.gg

Gemini CLI error

Usage limit reached for gemini-3.5-flash

Google fills in the model name, so the same message reads "Usage limit reached for gemini-2.5-flash." or "Usage limit reached for all Pro models." Under it the CLI prints the reset time and the stats, model and auth commands.

You used up the request allowance for that model.
Access resets at the time the message names.

Install now other errors

What it means

Gemini CLI keeps a request allowance per model, and the message names the model whose allowance is used up.

Run /stats to see each model's requests and usage left, with its own reset time. One model can be at 0% while another still has most of its quota.

The line is not always about your usage. In Gemini CLI 0.54.4 a 429 MODEL_CAPACITY_EXHAUSTED answer was not retried, and it showed as "Usage limit reached for gemini-3.5-flash" while /stats showed 1% to 8% used. The API's own words were "No capacity available for model gemini-3.5-flash on the server".

Some models are used by the CLI itself. A Gemini CLI maintainer closed one report by saying the CLI uses 2.5 Flash Lite for model routing, so exhausting that model blocks every prompt until its window resets, whatever model you picked.

Fix it

  1. Wait for the time after "Access resets at" in the message.
  2. Run /stats to see which model is out and when it resets.
  3. Run /model to switch to a model with quota left. If the CLI still blocks, the exhausted model may be one it uses for routing, as above.
  4. Run /auth to switch to a Gemini API key or Vertex AI, which bill per use instead of the plan window.
  5. If /stats shows quota left, update Gemini CLI. In the issue thread, 0.53.1 is named as the last release that retried the capacity answer.

Switch while you wait

A Gemini API key or a Vertex AI project bills per request, so the plan window no longer decides when you can work.

Or hand the task to another vendor you already pay for until the window resets.

The other model names

Usage limit reached for gemini 2.5 flash

Gemini CLI prints the model id with hyphens, as Usage limit reached for gemini-2.5-flash. The quota and the reset time work the same way. The stats command shows that model's usage left, and the model command moves you to one that has some.

Usage limit reached for all Pro models

This is the same check when every Pro model is out at once. The message adds a line for the stats, model and auth commands and offers Keep trying or Stop. Switch to a model with quota left, an API key or Vertex AI to get going again.

Usage limit reached for gemini-3.5-flash

The Gemini CLI quota for that model is used up, or the API answered that it has no capacity for the model. Wait for the time after Access resets at, switch models, or move to an API key or Vertex AI. If the stats command shows quota left, update the CLI.

With superbot

superbot is your own AI agent with a model of its own that runs on superbot credits, so it keeps working when a vendor's plan runs out.

It also puts your own API keys behind one base URL, with one caller key, in the OpenAI or Anthropic format.

Calls on your own keys spend no superbot credits, and you pay your provider as usual.

Gemini tools take an OpenAI or Anthropic base URL, so see the OPENAI_BASE_URL page.

Source: the Gemini CLI issues #28761, #26614 and #16578, and the quota and pricing doc. Read on 2026-11-10.