Sponsor of the Day:
Jerkmate
https://www.braintrust.dev/blog/model-updates
Bedrock, Vertex AI, and universal structured outputs - Blog - Braintrust
Full support for Bedrock, Vertex AI, and structured outputs in the AI proxy and playground.
vertex aistructured outputsblog braintrustbedrockuniversal
https://www.braintrust.dev/blog/brainstore-benchmarks
Brainstore makes AI observability at scale possible - Blog - Braintrust
Real-world benchmarks show Brainstore is up to 24x faster than competitors, making it possible to observe AI systems at production scale.
makes aiblog braintrustobservabilityscalepossible
https://www.braintrust.dev/blog/model-measurement
The 5 pillars of AI model performance - Blog - Braintrust
A framework for evaluating AI models across five dimensions, with Claude Opus 4.6 and GPT-5.3 Codex as case studies.
ai model performance5 pillarsblog braintrust
https://www.braintrust.dev/blog/measuring-what-matters
Measuring what matters: An intro to AI evals - Blog - Braintrust
Learn how to build effective evals for your AI products with datasets, tasks, and scores.
blog braintrustmeasuringmattersintroevals