Reddit user plots efficient frontier of open models
Defines efficiency as benchmark score over active parameters, using artificialanalysis.ai aggregate data. Only models on the Pareto frontier are included.
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Claude Code silently deletes local history older than 30 days
- Split sigmas speed up video reference in Stable Diffusion
- MiniMax tool turns anything into realistic human video
- MirroS' Code-as-World rewrites videos into executable MuJoCo programs
- PRAXIST research system boosts MLE-bench scores 44%