Skip to main content

Reasoning

Agentium supports extended thinking (reasoning) across all major model providers. When enabled, the model produces an internal chain-of-thought before generating the final answer — improving performance on complex math, logic, and multi-step problems.

Quick Start


Configuration

boolean
required
Enable or disable reasoning for this agent.
'low' | 'medium' | 'high'
Reasoning effort level. OpenAI only — maps to reasoning_effort parameter. Default: "medium".
number
Maximum tokens allocated for the thinking phase. Anthropic, Google, Vertex AI — maps to budget_tokens / thinkingBudget. Default: 10000.

Provider Support

Each provider maps ReasoningConfig to its native API:
When reasoning is enabled for OpenAI, temperature is ignored (the API does not allow both). For Anthropic, the temperature parameter is removed when thinking is active.

Output

When reasoning is enabled, RunOutput includes:

Streaming

During streaming, reasoning content is delivered as thinking chunks:

ReasoningConfig Type


Logging

When logLevel is "info" or higher, the agent logger displays:
  • Thinking content (truncated, italic, dimmed)
  • Reasoning token count with a brain icon in the usage summary

Accessing Thinking Content

When reasoning is enabled, the model’s internal thinking is available in RunOutput.thinking:
The thinking content is never shown to end users by default — it’s available for debugging, logging, or advanced use cases.

Reasoning with maxTokens

When reasoning is enabled on Anthropic, maxTokens must accommodate both thinking and response tokens. Agentium handles this automatically:
  • If maxTokens is set but too small for the thinking budget, it’s overridden to budgetTokens + 4096
  • For example: maxTokens: 1024 with budgetTokens: 2000 results in an effective max_tokens of 6096 sent to the Anthropic API
If you want precise control over max_tokens, set maxTokens to a value larger than budgetTokens + 1024.

Logging Reasoning Output

Enable logLevel: "debug" to see thinking content in the console:
At "info" level, only token usage summaries are logged. At "debug", the full thinking content is printed.