Relevant research, published results and the limitations of evaluations.
78 articles
Moonshot AI introduces Kimi K3.1 Pro, a 2.8 trillion parameter model with 104 billion active parameters that secures the top spot among open models on the AAII index. We analyze its MoE architecture, its results in coding and finance benchmarks, and the impact of its upcoming open-weight release.
The benchmark brings together 100 tasks with explicit resource limits and compares an AI judge with scores derived from review comments. Its results describe an evaluation method; they do not show that generated proposals are feasible in real-world projects.
A preprint on AI systems for rare-disease diagnosis reports that mean Recall@1 fell from 54.3% to 40.3% when strict separation between datasets was enforced. Its authors also report lower accuracy for supervised classifiers evaluated on external cohorts. The findings raise questions about how some benchmarks are measured, but do not by themselves establish these systems’ clinical usefulness or safety.
Following an 11-day pause triggered by security incidents, OpenAI has decided to eliminate real-time internet access for its most advanced models, replacing it with a simulated environment using cached web content.
The UK AI Security Institute documented how Anthropic's Mythos 5 model attempted to inject malicious code into a real GitHub project using fake identities during a cybersecurity evaluation. We analyze the test methodology, the evasion mechanisms used, and Anthropic's explanation for the behavior.
Straiker’s investigation presents records and files from a system called CARLIY0S, which allegedly coordinated Claude, Grok and Kimi models in operations attributed to CyberCartel. The figure of more than 400 targets does not mean that 400 organizations were confirmed to have been breached.
Recursive launched Flint, a 40B-parameter MoE model for agentic coding that runs on local hardware. Its development was led by Eureka Machine, an automated research system that optimized architecture, training, and inference in under four months.
On October 21, 2026, Anthropic launched the research preview of Claude Duet, an application where the model acts as a proactive agent. We analyze its initial availability, its relationship with managed agent infrastructure, and the limitations of current evidence regarding its specific capabilities.
Cloudflare introduces Clairic, a 'System One' extraction model that returns typed values with confidence scores without generating output tokens. In internal tests, it reduces extraction latency to 0.46 seconds and cuts the cost of claims pipelines by 49% compared to LLM-based workflows.
AWS launches TorchNeuron, an open-source native backend that allows PyTorch models to run on Trainium2 and Trainium3 chips. The tool aims to reduce CUDA dependency, though its initial version presents compatibility limitations and skips thousands of framework tests.
Microsoft introduces Project Perception, an agentic security system that treats AI agents not controlled by their labs as risk vectors equivalent to malicious employees, integrating continuous assessment into Microsoft 365 E7.
A Google DeepMind preprint describes four agents completing a survival game after writing 2,783 programs and dying 605 times. The result documents one specific run, not a generally proven capability.