Vals AI with a16z money is building an independent standard for AI benchmarks
· Source: original
Vals AI, backed by a16z, wants to become the gold standard for AI benchmarking
📊
The company is called Vals AI. The project is backed by the venture fund Andreessen Horowitz (a16z), and its stated goal is to become the "gold standard" for AI benchmarking.
What it does: conducts independent evaluations of AI models based on its own testing methodology. It positions itself as a neutral and trustworthy source of evaluations.
The mechanism is simple. The bet is not on vendors' marketing benchmarks, but on its own methodology: the same tasks, the same run conditions for all models. A vendor's benchmark shows a model in its own scenario; an independent methodology yields a comparable number across different models.
Why this matters: the wave of AI model releases keeps growing 📈, and each one comes with its own numbers from a blog post. An external evaluator with a fixed protocol provides comparable numbers across all models.
We broke down how the evaluation methodology works on techcrunch.com — published on September 19, 2026
🛠️ Want to put this into practice?
Inspired by the news — a selection of prompts for developers from our library:
- Code Review Agent
- Comprehensive Repository Analysis and Bug Fixing Framework
- Vacuum Arc Modeling under Transverse Magnetic Fields
🔗 The entire prompt library · "Programming" category
A ready-made product on the topic: Prompts for Content Creators — grab it and apply it right away.