LangChain Publishes Tutorial on Building an Agent Harness with Jev
Sydney Runkle of the LangChain team published "Building a Harness with Jev" on September 18, which both the official LangChain account and founder Harrison Chase reshared, making it one of the most-watched pieces in the LangChain ecosystem that week. The article frames an agent's execution as a loop — the LLM decides the next step, tools execute, the model evaluates the result, and the loop continues until the task is done — and walks through building an evaluation harness around that loop.
Confirmed
- Runkle notes that classification steps in many agent pipelines can be replaced by a small classification-specialized model like Jev, substantially reducing latency and cost.
- Jev specializes in constrained outputs; LangChain's official account said anyone confused about Jev should read the article carefully.
- Harrison Chase called Jev a good, cheap, fast classifier that, once brought into agent workflows, enables more workflow control primitives and adds controllability in LangGraph scenarios.
- matthewcanham called it one of the better breakdowns of how Jev works, viewing Jev as a new model optimized for decisions rather than text generation.
Why it matters
- The tutorial offers a hands-on path to building an agent harness, straight from official LangChain output.
- Offloading classification to a dedicated small model is a practical way to cut agent cost and latency, aligning with LangGraph's control-primitive philosophy and directly useful for production agent pipelines.
2026-09-18 ~ 2026-09-19 · 7 related posts
- Episode 1: TypeSafe exits stealth with decision model Jev and RLCD training method(2026-09-16, 86 posts)
- Episode 2: TypeSafe AI launches Jev, a dedicated evaluation model showing major speed and cost gains in tests(2026-09-16, 8 posts)
- Episode 3: Vercel fx to adopt Jev safety reviewer, up to 18x faster(2026-09-17, 3 posts)
- Episode 4: OpenJev Open-Source Clone Runs Jev-Style API on One RTX 3090(2026-09-17, 3 posts)
- Episode 5: TypeSafe AI Launches Jev, a Decision-Only Model That Outputs No Text(2026-09-17, 73 posts)
- Episode 6: Mystery decision model Jev sparks wave of community benchmarks(2026-09-18, 6 posts)
- Episode 7: TypeSafe AI's Jev model and RLCD method spark debate(2026-09-18, 3 posts)
- Episode 8: LangChain Publishes Tutorial on Building an Agent Harness with Jev(2026-09-18, 7 posts)
- Episode 9: Jev Ecosystem Explodes with Wave of Popular GitHub Projects(2026-09-18, 4 posts)
- Episode 10: Jev, the Decision-Only AI Model, Goes Viral for Speed and Cost(2026-09-18, 33 posts)
- Episode 11: TypeSafe Jev Decision Models Spark Debate on Breakthrough, Skepticism and Replication(2026-09-18, 7 posts)
- Episode 12: Jev Matches Classifiers at 100x Lower Cost(2026-09-19, 2 posts)
- Episode 13: Jev Positioned as LLM Router, Not a Replacement(2026-09-19, 2 posts)
- Episode 14: Developers Hail Jev as Top Value Model for Batch Labeling(2026-09-19, 2 posts)
- Episode 15: JevBench Launches: Jev 1.13 Holds Narrow Lead as Open-Source Rivals Close In(2026-09-19, 8 posts)
- Episode 16: TypeSafe AI's Decision Model Jev Explodes in Popularity, Setting Vercel AI Gateway Adoption Record(2026-09-19, 14 posts)
- Episode 17: omarsar0 Tests Jev as a New Primitive: Cheap Verifiers for Long-Running Agents(2026-09-19, 10 posts)
- Episode 18: Dev Tests 16 Models on Calibration: Jev Fast and Honest but Only 10th in Accuracy(2026-09-19, 8 posts)
- Episode 19: Independent Tests Confirm Jev Classifier's Accuracy at Fractional Cost(2026-09-19, 4 posts)
Primary sources
- [source] LangChain publishes guide on building an agent harness with Jev — LangChain · 2026-09-18
- LangGraph taps Jev as a cheap, fast classifier to add control-flow primitives to agent flows — hwchase17 · 2026-09-18
- LangChain's Jev targets constrained output for building agent harnesses — multiply_matrix · 2026-09-18
- Replace agent classification steps with optimized small models to cut latency and cost — LangChain · 2026-09-18
- Jev Is a New Model Optimized for Decisions, Not Text — Best Breakdown Yet — hwchase17 · 2026-09-19