Model Selection & API Keys
AgentJara supports multiple AI models from leading providers. Bring your own API keys for full control over costs and capabilities.
Supported Models
| Provider | Models | Best For |
|---|---|---|
| Anthropic | Claude Opus 4.6, Claude Sonnet 4.5, Claude Haiku 3.5 | Complex reasoning, writing, analysis, safety |
| OpenAI | GPT-5.3, GPT-5.2 Pro, GPT-5.2 Codex, GPT-4.5 | Advanced reasoning, coding, general tasks |
| OpenAI Codex | GPT-5.3 Codex | Powerful code generation |
| Gemini 3 Pro, Gemini 3 Flash, Gemini Flash Latest | Best value, fast, multilingual | |
| Meta | Llama 4 Scout, Llama 4 Maverick | Open-source, fast & efficient |
| DeepSeek | DeepSeek V3, DeepSeek R1 | Cost-effective, math & science reasoning |
| Groq | Groq Llama 3 70B | Ultra-fast inference |
| Cerebras | Cerebras Llama 3 70B | Ultra-fast hardware inference |
Adding API Keys
Go to Settings → API Keys and add your keys. AgentJara never stores your keys in plain text — they are encrypted at rest.
Per-Agent Model Selection
Each agent can use a different model. Choose based on the task: use a fast model like Gemini 3 Flash or Claude Haiku 3.5 for simple Q&A, and a powerful model like Claude Opus 4.6 or GPT-5.3 for complex reasoning.
Fallback Models
Configure a fallback model for each agent. If the primary model fails or is unavailable, the fallback model ensures uninterrupted service.