A curated firehose for the o1 reasoning explosion

What it does This is a curated “awesome” list that catalogs the research deluge around OpenAI o1 and LLM reasoning. It corrals papers, technical reports, blog posts, talks, open-source models, and training codebases into a single living document. The maintainer updates it continuously to track the frontier, which, given the pace of this field, is practically a part-time job.

The interesting bit The list doubles as a map of a hype cycle in real time. You can watch the narrative shift from early Twitter speculation and reverse-engineering blogs to actual open-source replications like DeepSeek-R1 and dedicated RL training frameworks. It is a bibliography of an arms race.

Key highlights

• Covers the full reasoning stack: news, blogs, talks, courses, models, and training code.
• Tracks major open-source challengers like DeepSeek-R1, Qwen3/QwQ, Kimi k1, and Skywork.
• Includes RL training codebases such as OpenRLHF, SimpleRL-reason, and Berkeley’s TinyZero.
• Maintains a deep paper list with technical reports from OpenAI, DeepSeek, NVIDIA, Mistral, and others.
• Curated by a maintainer who also publishes their own reasoning RL research (REINFORCE++, IcePop).

Caveats

• It is a link aggregator, not a framework or original research; the value is entirely in curation.
• The README is a long scroll with minimal categorization beyond broad headings; finding specifics requires Ctrl+F.
• Some sections (like Twitter screenshots) are just image dumps without context or summaries.

Verdict Worth bookmarking if you are actively working on LLM reasoning or trying to understand how o1-style models are trained. Skip it if you want executable code or a systematic literature review with analysis.