DeepSeek-V4-Flash-0731 is a new 304B parameter AI model with enhanced agentic capabilities, offering high performance at low cost.
Analysis of GPT 5.6's 72 configurations, mapping model choices and reasoning effort to performance and cost.
Microsoft announces MAI-Thinking-1 and MAI-Code-1-Flash LLMs, with details on parameters, licensing, and training data.
Microsoft announces MAI-Thinking-1 and MAI-Code-1-Flash LLMs, with low active parameters and claims of performance, but training data issues remain.
An overview of inference-time scaling methods for improving LLM reasoning, categorizing techniques and highlighting recent research.
An overview of inference-time scaling methods for improving LLM reasoning, categorizing techniques like chain-of-thought and self-consistency.
Explores Abstraction of Thought (AoT), a structured reasoning method that uses multiple abstraction levels to improve AI reasoning beyond linear Chain-of-Thought approaches.
OpenAI releases GPT-5.1 API with new reasoning modes, adaptive reasoning, extended prompt caching, and new built-in tools for developers.
Explores how increasing 'thinking time' and Chain-of-Thought reasoning improves AI model performance, drawing parallels to human psychology.
An introduction to reasoning in Large Language Models, covering key concepts like chain-of-thought and methods to improve LLM reasoning abilities.
A technical guide on fine-tuning IBM's Granite3.1 AI model using Guided Reward Policy Optimization (GRPO) to enhance its reasoning capabilities.
A tutorial on reproducing DeepSeek R1's RL 'aha moment' using Group Relative Policy Optimization (GRPO) to train a model on the Countdown numbers game.
