Qwen3.8-27B ModelScope Model CardProvider report

Name: Qwen3.8-27B ModelScope Model Card
Published: 2026-08-15

2026-08-15Based on the master branch of the Qwen/Qwen3.8-27B ModelScope model card. The ModelScope API was last updated at 2026-08-14T23:00:10+08:00; the card was captured on August 15, 2026, and its content is pinned by README SHA-256 be2ef1984e6f654eaaf410f56d52aac962d85654fac259803bede66e81cad041. This snapshot includes all 12 source rows from Text Performance and all 13 source rows from VL Performance. It splits the compound cells for Agents' Last Exam, ClawEval-MM, MathVision, BabyVision, and CharXiv (RQ), producing 30 metrics and 114 available scores. All five comparison-model columns are preserved as reported; '--' cells do not create scores. QwenSWEBench, CoWorkBench, and RecreationBench are Qwen internal benchmarks.

Back to data sources

Qwen3.8-27B ModelScope Model CardProvider report

View source

2026-08-15Based on the master branch of the Qwen/Qwen3.8-27B ModelScope model card. The ModelScope API was last updated at 2026-08-14T23:00:10+08:00; the card was captured on August 15, 2026, and its content is pinned by README SHA-256 be2ef1984e6f654eaaf410f56d52aac962d85654fac259803bede66e81cad041. This snapshot includes all 12 source rows from Text Performance and all 13 source rows from VL Performance. It splits the compound cells for Agents' Last Exam, ClawEval-MM, MathVision, BabyVision, and CharXiv (RQ), producing 30 metrics and 114 available scores. All five comparison-model columns are preserved as reported; '--' cells do not create scores. QwenSWEBench, CoWorkBench, and RecreationBench are Qwen internal benchmarks.