large reasoning model (LRM)

A large reasoning model (LRM) is a language model optimized for multi-step problem-solving that allocates extra computation and uses structured intermediate steps during inference to plan, verify, and refine its answers.

LRM capabilities are built on top of standard LLMs through training and inference techniques, including some of the following:

• Reinforcement learning on reasoning traces
• Explicit reasoning tokens for test-time thinking
• Search or self-consistency to improve correctness
• A reasoning effort or thinking budget control that sets how much the model deliberates before answering

In current practice, this reasoning is usually a tunable mode on a general-purpose frontier model rather than a property of a separate model class. The same model can answer directly on easy inputs and deliberate at length when you raise the effort setting.

Many LRMs also use external tools, such as code execution, to enhance their reasoning.

Related Resources

Course

Building Type-Safe LLM Agents With Pydantic AI

Build type-safe LLM agents in Python with Pydantic AI using structured outputs, function calling, and dependency injection.

intermediate ai