One of the most significant AI advances since the last Radar is the breakthrough and proliferation of reasoning models. Also marketed as "thinking models," these models have achieved top human-level performance in like frontier mathematics and coding.
Reasoning models are usually trained through reinforcement learning or supervised fine-tuning, enhancing capabilities such as step-by-step thinking (), exploring alternatives () and . Examples include OpenAI¡¯s /, DeepSeek R1 and . However, these models should be seen as a distinct category of LLMs rather than simply more advanced versions.
This increased capability comes at a cost. Reasoning models require longer response time and higher token consumption, leading us to jokingly call them "Slower AI" (as if current AI wasn¡¯t slow enough). Not all tasks justify this trade-off. For simpler tasks like text summarization, content generation or fast-response chatbots, general-purpose LLMs remain the better choice. We advise using reasoning models in STEM fields, complex problem-solving and decision-making ¡ª for example, when using LLMs as judges or improving explainability through explicit CoT outputs. At the time of writing, Claude 3.7 Sonnet, a hybrid reasoning model, had just been , hinting at a possible fusion between traditional LLMs and reasoning models.
