Recall
Pinned Loading
Repositories
- llm-tournament-benchmark Public
Pairwise LLM benchmark, August 2025: 45 scenarios across 8 skills, 51 models, Swiss tournaments, LLM judges and OpenSkill ratings. Results in this repo are quick runs of 4 models. To rerun 2 summarization scenarios, download arXiv 2412.14498 and 2504.17028 into documents/. Their licenses do not let this repo include them.
- electric-pelican Public archive
A platform for evaluating and intelligently routing AI models using real head-to-head performance data and hierarchical capability scores
- air-coding-arena Public archive
Reference application tracking Linear coding-agent competitions, GitHub pull requests, and OpenSkill leaderboards.
- thwip Public archive
Reference implementation of signed local CI proofs and an ordered GitHub App merge queue.
- astrochicken Public archive
Reference platform for orchestrating coding-agent workflows and local software factories
- memracer Public archive
Reference prototype for mediated human and coding-agent coordination over MCP, with durable task queues and a dashboard.
- skills-are-all-you-need Public archive
- survival-receipts Public archive