WhitzardNUWA Lab ↗
NUWANUWA INDEX

Search NUWA Index

Research · Benchmarks · Findings · Perspectives

中文
← Assessments and reports

DEVELOPER DISCLOSURE · Z.ai

GLM-5: from Vibe Coding to Agentic Engineering

Developer documentation covering GLM-5.

Published Not verified

Sources ↓

Developer-authored disclosure. Selected source sections were read; numerical results and comparability have not been audited. NUWA has not independently replicated these experiments.

Models covered

Report coverage

SectionSource locationTreatment
3.4 General RL; agentic trainingOriginal source → 3.4 General RL; agentic trainingContextual evidence

Reward-model robustness and training-time reward hacking are discussed; there is no dedicated catastrophic-risk assessment in the reviewed report.

References

  1. [1] GLM-5: from Vibe Coding to Agentic Engineering