Benchmark Almanac

Preferred Networks

We track 2 benchmarks that Preferred Networks made or helped make.

0 in use 2 with public code 1 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
JFBench Models answer Japanese prompts with multiple verifiable constraints. A response passes only if it satisfies every constraint. Chat and instructions in many languages Mar 2026 48.9% Not in cards
BBH-ja Models solve Japanese translations of BIG-Bench Hard reasoning problems. Reasoning in many languages Jul 2025 Not in cards