Models
UniKey exposes leading AI model families through a unified gateway. Use /v1/models to inspect the models currently available to your account before hard-coding model IDs in production.
Common model IDs
The model catalog may change as upstream providers release, rename, or retire models. The following IDs reflect the model families used in the current API examples:
| Family | Example IDs |
|---|---|
| OpenAI-compatible | gpt-5.2, gpt-5.4, gpt-5.4-mini, gpt-5.4-nano, gpt-5.5, gpt-5.5-instant |
| Anthropic-compatible | claude-sonnet-4-6, claude-opus-4-6, claude-haiku-4-5 |
| Gemini | gemini-3.5-flash, gemini-3-flash |
| DeepSeek | deepseek-v3.2 |
| Other providers | kimi-k2.5, glm-5.1, minimax-m3, minimax-m2.7 |
Alias handling
Some catalog entries may use dot-version aliases such as claude-sonnet-4.6. API requests and Claude Code configuration should prefer the hyphenated alias style used in these docs, such as claude-sonnet-4-6.
Choosing a model
- Use larger reasoning models for complex planning, coding, and analysis.
- Use smaller or flash models for latency-sensitive, high-volume tasks.
- Use Claude-compatible models when a client expects the Anthropic Messages API.
- Use
stream: truefor long responses where progressive output improves user experience.
Production guidance
Do not assume every account has access to every model. Always handle 403, 404, rate-limit, and insufficient-credit responses gracefully, and keep a fallback model configured for critical workflows.