Post · Benchmarks & research
Jev as an LLM Judge
Yuchen Jin's take on Jev for LLM-as-a-judge work: 20-200x faster and 40-400x cheaper, with a screenshot of the model picking which candidate is AGI…
Open x.com ↗Jev has spoken.
It picked which model is AGI.
20–200x faster. 40–400x cheaper. This could make things like LLM-as-a-judge insanely fast and nearly free.
(I tried a bunch of prompts and still didn’t burn through $0.10.)
How builders describe it
Yuchen Jin's take on Jev for LLM-as-a-judge work: 20-200x faster and 40-400x cheaper, with a screenshot of the model picking which candidate is AGI after a prompt sweep under $0.10.
The decision Jev makes
Benchmark questions with known answers, to check accuracy and confidence.
Where it fits
Head-to-head tests, calibration studies and independent research into how well Jev decides, how fast, and at what cost. All 355 benchmarks & research projects →
Related projects
Benchmarks & researchjev-browsecompResearch harness comparing cheap TypeSafe Jev typed decisions (Choice/Noul act|review|abstain over stable doc ids) to full-LLM and Recursive Language…GitHub · jjd-lab
Benchmarks & researchjev-ticket-triageReproducible support-ticket triage eval: TypeSafe Jev vs Together.ai LLMs on accuracy/cost/latency/confidence (WIP).GitHub · beese54Benchmarks & researchtev1Together AI's open recipe and weights for a Jev-inspired decision model fine-tuned on Qwen3.5-4B, with the full data pipeline, training config, and…GitHub · ★ 141 · Together AI
Benchmarks & researchJev-StyleSmall calibrated decision models you run locally, with a systemone-compatible server, agent skills, Claude Code guard, and MCP tools—weights on…GitHub · ★ 2 · lawrence3699
Benchmarks & researchJevletFrom-scratch research reconstruction of a Jev-like System One decision model (typed Noul/Choice/Score → probabilities, no text generation) plus a…GitHub · ★ 1 · NAME0x0
Benchmarks & research50 financial jobs, one modelThe Fintech Builder runs Jev on fifty financial jobs and shows where it works and where it does not.YouTube · youtube.com/@TheFintechBuild
Benchmarks & researchjev-ticket-triageReproducible support-ticket triage eval: TypeSafe Jev vs Together.ai LLMs on accuracy/cost/latency/confidence (WIP).GitHub · beese54Benchmarks & researchtev1Together AI's open recipe and weights for a Jev-inspired decision model fine-tuned on Qwen3.5-4B, with the full data pipeline, training config, and…GitHub · ★ 141 · Together AIBenchmarks & researchJev-StyleSmall calibrated decision models you run locally, with a systemone-compatible server, agent skills, Claude Code guard, and MCP tools—weights on…GitHub · ★ 2 · lawrence3699Benchmarks & researchJevletFrom-scratch research reconstruction of a Jev-like System One decision model (typed Noul/Choice/Score → probabilities, no text generation) plus a…GitHub · ★ 1 · NAME0x0Benchmarks & research50 financial jobs, one modelThe Fintech Builder runs Jev on fifty financial jobs and shows where it works and where it does not.YouTube · youtube.com/@TheFintechBuild