Models for testing and CI: fast and cheap
In development and CI you want fast replies and predictable cost, not the most expensive model. This collection gathers flash/lite and mini options for prompt smoke tests, automated evals, and team sandboxes.
Models in this collection
| Model | Type | Context | Price |
|---|---|---|---|
 Gemini 2.5 Flash-LiteGoogle | Chat | 1M tokens | $0.109/M in · $0.436/M out |
 DeepSeek V4 FlashDeepSeek | Chat | 1M tokens | $0.1526/M in · $0.3052/M out |
 Ministral 8BMistral | Chat | 128k tokens | $0.109/M in · $0.109/M out |
 GPT-4.1 mini (legacy)OpenAI | Chat | 1M tokens | $0.436/M in · $1.7441/M out |
| Gemini 3.1 Flash-LiteGoogle | Chat | 1M tokens | $0.2725/M in · $1.6351/M out |
 Claude Haiku 4.5Anthropic | Chat | 200k tokens | $1.0901/M in · $5.4503/M out |
| Gemini 2.5 FlashGoogle | Chat | 1M tokens | $0.327/M in · $2.7251/M out |
Why these models
Low latency + low price + enough quality for asserts. We skip expensive flagships: in CI the model should fail cheaply and quickly when a prompt or JSON contract breaks.
Use them with Geek Hub
One OpenAI-compatible base URL and API key. Swap any model id from this list without rewriting your client.
Get an API keyFAQ
- Which model should I use in CI?
- Gemini Flash Lite or DeepSeek Flash usually cover smoke tests. If you validate structured outputs or tool calling, GPT-4.1-mini or Haiku are more stable.
- Should test and production use the same model?
- Not necessarily. Test with a cheap one and pin production via env. Keep the contract (JSON, tools) the same.
- How do I cap spend on testing?
- Use a Geek Hub API key just for CI with a low budget, and models from this list. Swap the model id via environment variable.