Reasoning Model
Also known as: Thinking model, o1-style model
An LLM trained to reason at length internally before answering. Slower and more expensive, but much better on hard problems.
An LLM trained (often via RL) to produce long internal chains of reasoning before its final answer. Examples: OpenAI o1/o3, DeepSeek-R1, Gemini Thinking. Trade higher latency and cost for much stronger performance on math, code, and multi-step problems.
In practice
The 2024-2025 paradigm shift. Senior interviews probe when reasoning models are worth the latency/cost and how their inference budget is controlled.
How it compares
Reasoning models bake CoT into RL training; CoT alone is just a prompting technique on any model.
Related topics
Related terms
RLHF (Reinforcement Learning from Human Feedback)
Train a reward model from human preference pairs, then RL-fine-tune the LLM against that reward.
Hallucination
When a model confidently makes up something that isn't true.
LLM Evaluation
Measuring whether an LLM does what you want, beyond "looks fine to me".
DPO (Direct Preference Optimization)
Skip the reward model and PPO: fine-tune directly on preferred-vs-rejected response pairs.
Perplexity
Exp(average cross-entropy) on held-out text; lower means the model is less surprised by real data.
Pre-training
Self-supervised next-token training on a huge unlabeled corpus, producing the base model.