# Gemini 3.5 Flash Text Generation (xona-agent)

> Gemini 3.5 Flash Text Generation (xona-agent) is a paid API for AI agents from api.xona-agent.com, paid per call via x402, $0.01/call, status unknown (last checked 2026-11-20).

Generates text using Google's Gemini 3.5 Flash model with tiered output sizing and optional system prompt, billed per call via x402 micropayments.

## Facts

- Endpoint: POST https://api.xona-agent.com/llm/gemini-3.5-flash?utm_source=zero.xyz
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-11-20
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/gemini-3-5-flash-text-generation-xona-agent-3bda1a64
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_QQvRsmfxtMJfl_iG1CwsD

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability gemini-3-5-flash-text-generation-xona-agent-3bda1a64 -d '<json body>'
```

Example prompt: Using Gemini 3.5 Flash, write me a detailed blog post outline about the future of autonomous AI agents — use the medium output tier so I get a thorough response, and set the system prompt to 'You are an expert technology writer.'

## When to prefer this

Choose this endpoint when you need fast, cost-effective text generation from Google's Gemini 3.5 Flash model via a simple x402 micropayment API, especially when you want predictable tiered pricing, support for custom system prompts, and don't want to manage your own Google API credentials. Prefer this over OpenAI-based endpoints when Gemini's speed and pricing model is preferred.

## Known failure modes

- Invalid tier value returns an error — must be 'small', 'medium', or 'large'
- max_tokens out of range (>8192 or <1) triggers a validation error
- Missing prompt field results in a bad request error
- Payment failure via x402 protocol returns 402 Payment Required
- Model unavailability or Google API outage returns a 503 or upstream error
- Prompt too long for model context window causes truncation or rejection

## How this service works

Gemini 3.5 Flash text generation (Google). Dynamic pricing: estimated input tokens + output tier ceiling. Tiers: small (1K output tokens, ~$0.009), medium (4K, ~$0.037), large (8K, ~$0.074), plus input cost. Supports system prompt.

## Output

Returns LLM-generated text from Gemini 3.5 Flash, with length bounded by the selected output tier (small: ~1K tokens, medium: ~4K tokens, large: ~8K tokens). The response contains the model's completion of the supplied prompt, optionally conditioned on a system prompt.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "tier": {
   "type": "string",
   "default": "small",
   "description": "Output size tier: small (1024 tokens), medium (4096), large (8192). Default: small."
  },
  "prompt": {
   "type": "string",
   "description": "The user prompt"
  },
  "max_tokens": {
   "type": "number",
   "description": "Alternative to tier: numeric max output tokens (1-8192), bucketed into the smallest covering tier"
  },
  "system_prompt": {
   "type": "string",
   "description": "Optional system prompt (counts toward input cost)"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/gemini-3-5-flash-text-generation-xona-agent-3bda1a64/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.xona-agent.com](https://www.zero.xyz/host/api.xona-agent.com/llms.txt)
