Long Context Pre-Training with Lighthouse Attention

When large language models (LLMs) hallucinate, developers typically assume the model lacks the required facts. Engineering teams diagnose the error as missing knowledge. The standard response is to increase model size, expand training data, or build complex retrieval architectures.A new study by res

Sep 1, 2026

Earlier this year, we ran a pilot giving external researchers access to aggregate, real-world Claude usage data. Three research groups designed their own studies for Anthropic Insights, our privacy-preserving analysis tool. In this post, we share high-level results from those studies and what we learned running this pilot.

Aug 27, 2026

Consider an AI agent tasked with a complex enterprise workflow like migrating massive batches of customer records from a legacy CRM to a cloud database. The agent cannot rely solely on its internal context window for a job spanning hours and depends on the runtime layer, aka the harness. This harnes

Aug 28, 2026

Chroma research finds that LLM performance varies significantly as input length increases — a phenomenon they call 'context rot.' Across an evaluation of 18 LLMs including GPT-4.1, Claude 4, Gemini 2.5, and Qwen3, the study shows that models exhibit degraded retrieval and reasoning ability as context windows grow longer, challenging the common assumption that LLMs process context uniformly.

Jul 13, 2025