22M local model beats JEV 93% vs 80% on Banking77 in 8ms on CPU

Prompt Engineering · youtube · 2026-09-20

The Prompt Engineering channel benchmarks JEV—touted as a brand-new model category—against hype and history:

• Single-pass classifiers date to 2018 and zero-shot label pipelines to 2019, so JEV isn't a new concept.
• Across four tasks vs three established approaches, a 22M-parameter local model scored 93% on Banking77 vs JEV's 80%, running in 8ms on CPU for free.
• JEV's real edge: dynamic instruction following, something the older approaches can't match.
• The video also covers the 'System One models' framing and hybrid architectures.

Original post →

More from Models

• XGEN debuts Generative World Simulation: JING model tops WBench Full split — hey_abusiddik · 2026-09-20
• Code-only heuristic policies can beat frontier models on Craftax, evals researcher says — JoshPurtell · 2026-09-20
• Codex usage reset now live for all, big OpenAI release teased for Tuesday — kimmonismus · 2026-09-20
• Jev beats GPT-5.6 Luna on PR review: 1.93x faster at $0.0014 per run — aniketmaurya · 2026-09-20
• Fruit fly connectome chess model beats Jev 4-1 in 10 games, with a playable demo site — maximelabonne · 2026-09-20
• Bonsai 2 27B safety guardrails reportedly cut SWE-bench and Terminal-bench scores by ~20 points — julianharris · 2026-09-20

Models channel →