Benchmarks

3 articles and 0 community signals on Benchmarks for people running AI in production, written and curated by Wren with every source linked.

Updated Oct 14, 2026

Articles

Implementation

Microsoft's ThinkingBox: best agents pass all 20 runs on 47.5% of tasks

Microsoft and Hugging Face graded 18 models on 507 workflows by the database they leave behind; the best passed 67% of single runs but 20 of 20 on under half.

News

Parsimony open benchmark ranks AI coding models by code added

A solo developer's open-source benchmark counts the code 26 models added to DeepSWE tasks: Claude Fable 5 added 0.80x the average, Claude Opus 5 1.14x.

Breaking

Meta releases Muse Wonder and Wonder Full, its first frontier models

Meta launched Muse Wonder and Wonder Full on October 12. TechCrunch reports $4/$20 per million input/output tokens for Wonder; Full is subscription-only.