A model guide for the GPT-6 family
OpenAI 2 weeks ago 6
A model guide explains how startups can choose GPT-6 models, adjust reasoning effort, and improve prompts and tool coordination for production workflows.
Large Language Models
443 summarised stories about Large Language Models, each linking back to the original source. Browse all topics →
Key players
Friday, 2 October 2026
Fine-tune a search agent with multi-turn RL on Amazon SageMaker AI
Amazon Web Services 2 weeks ago 29
Amazon SageMaker AI fine-tuned a Qwen3.6-27B search agent using multi-turn reinforcement learning to improve how it searches across multiple interaction rounds. The failure rate on BrowseComp-Plus fell from 22.89% to 0.68%. Multi-turn RL raised retrieval quality on three of four held-out benchmarks and made the agent substantially more reliable by reducing task-completion errors within the turn and token limits.
Open-sourcing AstaBrief, the fast report-generation model in Asta
Allen Institute (AI2) 2 weeks ago 13 ● 2 sources
Open-sourcing AstaBrief, a fast report-generation model built in Asta for producing cited scientific reports from research questions and retrieved excerpts.
Don’t be fooled—LLMs don’t reason
MIT Technology Review 2 weeks ago 21
Thore Graepel argues that AlphaGo’s famous move 37 reflected true reasoning from an explicit search over future possibilities, not “intuition” alone, and contrasts this with how today’s large language models generate answers token by token. He says LLM reasoning attempts via techniques like “chain of thought” still rely on the same next-token prediction process and often lack an explicit epistemic state, citing move 37 as a case where the policy ranked it as about a 1 in 10,000 chance. He concludes that trustworthy AI systems need a new machine-reasoning approach modeled on AlphaGo’s architecture, including an auditable state of beliefs that gets updated with evidence rather than post-hoc rationalizations.
AWS Strands Labs Releases Strands Decider 2B: An Open Source Decision Model That Picks Options in About 115 ms
MarkTechPost 2 weeks ago 48 ● 5 sources
AWS Strands Labs released Strands Decider 2B, an open-source decision model that reads a state and typed questions and outputs an option choice, a yes/no probability, or a rubric score (no text generation). It has 1.9 billion parameters and runs with a reported 115 ms median latency on an RTX 3090. The release includes Apache-2.0 weights on Hugging Face plus a pip-installable CLI and local HTTP server, with self-hosted deployment supported but no hosted inference providers yet.
Academia is for Ambition — Alex Zhang, MIT
Latent Space 2 weeks ago 33
Alex Zhang of MIT discussed research on recursive language models and “harnesses” as compositional systems in an interview/podcast focused on AI tooling and agent architectures. The first 30 Latent Space subscribers get a 30% off code for new tickets (no refunds), with the event set for 2 weeks later. The episode shifts from past GPU-kernel communities to how AI-written kernels and multi-agent swarms raise capability and verification questions that shape future research bets.