liveBook · Manning

Reasoning model

As explained in LLM Customization and Fine-Tuning

A model optimized to perform multi-step reasoning tasks, especially those requiring structured deliberation.

Reasoning models tackle complex, multi-step tasks by producing intermediate reasoning before an answer. They can reduce the need for chain-of-thought examples in prompts, but generally require more processing time and longer context windows. section 4.1.1 section 5.2.2

Choose and evaluate them with the task and operating cost in view:

  • For math, code, or planning, prefer a reasoning-specialist model as a distillation teacher so the student can learn reasoning traces, not only final answers. section 7.1.2
  • Allow longer output lengths, filter for coherent reasoning, and evaluate intermediate steps as well as final answers. section 7.1.2
  • Recheck few-shot baselines on current frontier models; reasoning models may need fewer examples, while added examples can increase inference compute. section 4.1.1

Selecting a teacher

For a math or planning distillation task, a DeepSeek-R1-style reasoning specialist may be a better teacher than a general-purpose model. Check its accuracy on curated examples before generating training data. section 7.1.2

How LLM Customization and Fine-Tuning covers it

How deep it goes
Defines intermediate reasoning and gives selection and generation constraints, but does not detail model training.
Examples
Names OpenAI o-series planners and DeepSeek-R1-style teachers for math, code, and planning.
What you can do after
Apply the approach by selecting a reasoning-specialist teacher, budgeting longer outputs, filtering reasoning traces, and evaluating intermediate steps.
How it connects
Connects reasoning models to chain-of-thought prompting, in-context learning, knowledge distillation, and reinforcement learning from verifiable rewards.

Also explained in

Compare how 7 books explain reasoning model →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of LLM Customization and Fine-Tuning. The section links point to the parts of the book it draws on.