Home /Open models and replicas /typesafe-ai-benchmark
typesafe-ai-benchmark
LLM gateway that mimics the TypeSafe response shape, useful as a stand-in while you wait for a key.
From the readme
inference-benchmarks Benchmarks for model × version × engine. A score is never just the model's. The same weights score differently at another quant, or in another engine. So every result here names the exact checkpoint, quant and engine that produced it, and one harness measures them all the same way. Hackers in the Loop · runs on recipes from …
Details
- Section
- Open models and replicas
- owner
- iammrduncan
- stars
- 38
- forks
- 5
- pushed
- 2026-10-18
- Language
- TypeScript
- License
- MIT
What it is
- Form
- Proxy or service
- Host agent
- Standalone
- Audience
- Researchers
Maturity
Docs
reports measured numbers