BenchmarkList Tracks 2.4k+ AI Benchmarks
davidthesong · reddit · 2026-07-16
BenchmarkList is a public platform aggregating 2.4k+ AI benchmarks, models, and capability results. It offers continuously updated evaluation streams, model comparisons, and in-depth research pages like the "AI progress on human work" dashboard and visualizations.\n\nThe author notes that current AI capability evaluations are scattered across papers, GitHub, model cards, websites, and tweets. This unified entry point aims to help researchers systematically understand what models "can and cannot do."
Related event: BenchmarkList Launches with Over 2,400 AI Benchmarks(4 posts)→
More from Research
- Goodfire commits $10M to launch a 'Human Genome Project for AI' interpretability effort — mathildepapillo · 2026-10-27
- Researchers call for a Human Genome Project for AI to fully understand neural networks — mathildepapillo · 2026-10-27
- Jon Barron: every softmax and cross-entropy implementation has two subtle numerical bugs — jon_barron · 2026-10-27
- New theory paper explains what Anthropic's Jacobian lens actually detects in LLMs — wgilpin0 · 2026-10-27
- Interpretability researcher joining Goodfire AI post-PhD, calls for speeding up the field — mathildepapillo · 2026-10-27
- EdgeRock-4B holds KV cache at 72 MiB across 256K context, 99.1% smaller than Qwen3.5-4B — alifcoder · 2026-10-27