#Research Articles

10 articles
Filter by importance: AI Rating: High AI Rating: Medium AI Rating: Low
Filter by specialty: Chat & General Models Image Generation Video Generation Voice & Music Coding Search & Agents
Perplexity News 09/30/2026 Read AI Rating: Medium

Perplexity Releases a Contextual Embedding Model That Retrieves Evidence, Not Just Answers

Perplexity published research on September 30, 2026 introducing pplx-embed-v2-context-9b-preview, a 9B-parameter embedding model trained to retrieve supporting context alongside answers, plus Context-bench, a new benchmark for evaluating that capability.
Sakana AI News 09/28/2026 Read AI Rating: Medium

Sakana AI's SAIL Boosts Robot Trajectory Success From 25% to 73% via Test-Time Scaling

Sakana AI and the University of Tokyo unveiled SAIL, a method that uses iterative refinement and search at inference time to make vision-language models generate far more reliable robot trajectories, accepted to IROS 2026.
Sakana AI News 09/24/2026 Read AI Rating: High

Jürgen Schmidhuber Joins Sakana AI as Chief Scientific Advisor

AI pioneer Jürgen Schmidhuber has joined Sakana AI as Chief Scientific Advisor, contributing to its Recursive Self-Improvement Lab while pushing the company toward Physical AI and world models.
Sakana AI News 09/18/2026 Read AI Rating: Medium

Sakana AI Launches Frontier Intelligence Group to Explore Beyond-Transformer AI

Sakana AI unveiled its Frontier Intelligence Group (FIG), a research collective pursuing speculative, biologically inspired alternatives to today's transformer models, alongside an NVIDIA-backed sparse compute format called TwELL.
Devin News 09/09/2026 Read AI Rating: High

Cognition Says Devin Helped Factor RSA-260, Cutting Cost 10x With a New GPU Lattice Siever

Cognition published research showing how a researcher running concurrent Devin sessions built a GPU-based lattice siever that factored the 260-digit RSA-260 semiprime for about $400,000, roughly a tenth of the prior state-of-the-art cost.
ChatGPT News 09/08/2026 Read AI Rating: Medium

OpenAI Funds $5 Million in Independent Research on AI and Teen Development

OpenAI opened a $5 million grant program, with up to $1 million per project, funding independent researchers to study how generative AI affects the development, wellbeing, and social lives of 13-to-17-year-olds.
Claude News 08/27/2026 Read AI Rating: Medium

Anthropic Opens 10,000 Free Claude Team Seats for Scientists

Anthropic is giving academic and nonprofit researchers free or discounted access to Claude Team, widening its AI for Science program beyond biology, and offering up to $50,000 in compute credits per project.
ChatGPT News 08/17/2026 Read AI Rating: Medium

New policy ideas for the Intelligence Age

OpenAI is granting $1 million in direct funding plus up to $1 million in API credits to 14 independent organizations across five countries to prototype practical AI policy ideas covering economic opportunity and societal resilience.
ChatGPT News 07/28/2026 Read AI Rating: Medium

Scientific Computing in the Age of Agentic AI: OpenAI Publishes Field Report

An OpenAI field report shows researchers using coding agents like Codex and Claude Code to modernize eight scientific computing projects, including genomics tools, while highlighting that expert verification remains essential.