Skip to content
View original post on X: ModelScopeOfficial· 40/100AI score40/100

JEV-27B-VL ranks first on Decision Index Vision with calibrated visual decisions

AISummary

AutoTrust Labs releases JEV-27B-VL, which ranks first among 20 models on the independent Decision Index Vision board with 77.8% accuracy across 13,695 examples.

Its confidence scores closely track real accuracy, with an ECE of 0.019, and it supports image-and-text prompts up to 256K tokens.

The model is released with open weights under Apache 2.0, and its decision adapter, trained only on text, exceeds the 397B Qwen3.5 reference by 6.4 points on the board's full score in zero-shot vision.

Post on XView on X
ModelScopeVerified on X
@ModelScope2022

JEV-27B-VL brings accurate, calibrated visual decisions to long-context workflows.
🤖 https://modelscope.ai/models/AutoTrust-Labs/JEV-27B-VL
📄 https://modelscope.ai/papers/2512.16899

🏆 Ranks #1 among 20 models on the independent Decision Index Vision board, with 77.8% accuracy across 13,695 examples.
🎯 Its confidence scores closely track real accuracy, achieving an ECE of 0.019 for reliable threshold-based decisions.
📚 System 1 supports image-and-text prompts up to 256K tokens, with yes/no, 0–5 scoring, or selection from as many as 256 options.
⚡ Each System 1 query returns a calibrated probability for every option in one forward pass, while System 2 handles step-by-step multimodal reasoning.
👁️ The decision adapter was trained only on text, yet its zero-shot vision performance exceeds the 397B Qwen3.5 reference by 6.4 points on the board’s full score.
📜 Open weights under Apache 2.0.

Source: ModelScope · x.comPublished · added here

Topics

Multimodal AIOpen-source ecosystemModel releasesEvals & benchmarks