Claude Research

active

Multi-agent research system — Opus 4 lead + Sonnet 4 subagents. 90.2% improvement over single-agent (self-reported). 1M context, Agent SDK for custom pipelines.

Score 38

Where it wins

Multi-agent architecture: Opus 4 lead + Sonnet 4 subagents running in parallel

90.2% improvement over single-agent on internal eval (methodology published)

1M context window (Opus 4.6, Feb 2026)

Agent SDK enables custom research pipelines

Strong at long-form synthesis and citation quality

Where to be skeptical

90.2% improvement claim is self-reported — no independent benchmark validation

No dedicated HN thread for Claude research specifically

Lower independent community signal than Perplexity (368 pts), NotebookLM (907 pts), Tongyi (365 pts)

No standalone research product — embedded in Claude Pro/Max

Editorial verdict

#7 in research. Architecturally distinct multi-agent approach is theoretically superior, but 90.2% improvement claim is self-reported with no independent validation. No dedicated HN thread for Claude research. Promising but needs third-party benchmarks to justify a higher ranking.

Source

Docs: docs.anthropic.com

Found via SkillPack? ★ Star us on GitHub

Research

#07of 12

Long-form synthesis, custom research agent development

GPT Researcher

Open-source autonomous deep research agent. CMU DeepResearchGym #1 on citation quality, report quality, info coverage. 25.8K stars, 15.9K weekly PyPI downloads. Apache 2.0.

Tongyi DeepResearch

First fully open-source deep research agent matching closed-source leaders on benchmarks. HLE 32.9 (exceeds OpenAI's 26.6), 30.5B params / 3.3B active (MoE), runs locally. 18.5K stars. Apache 2.0.

STORM (Stanford)

Stanford's LLM-powered knowledge curation system. Generates Wikipedia-style articles with citations in ~3 min. 28K stars, 84.8% citation recall / 85.2% precision (peer-reviewed). MIT license.

Perplexity Deep Research

Research-first search engine with inline citations. Fastest deep research (15-30s), 93.9% SimpleQA accuracy, 50+ sources per report. $20/mo Pro.

Public evidence

strongSelf-reported2025

Multi-agent architecture: 90.2% improvement over single-agent

Architecturally distinct approach. Self-reported but methodology is published in detail.

Engineering blog with published methodologyAnthropic engineering team

moderateSelf-reported2026-02-05

Opus 4.6 with 1M context window (Feb 5, 2026)

1M context competitive with Gemini. Agent SDK enables custom research pipelines.

Official announcementAnthropic

moderate2026-03-20

No dedicated HN thread for Claude research

Lower independent community signal compared to competitors. Most Claude discussion is general, not research-specific.

HN Algolia search — no resultsN/A

Raw GitHub source

GitHub README could not be fetched right now.