NVIDIA adds Legal Agent Bench to NeMo Gym with 1,749 tasks and public office-file skills
NVIDIAAI · x · 2026-08-04
NVIDIA’s NeMo Gym now includes Legal Agent Bench, a 1,749-task benchmark for evaluating and improving agents in legal workflows.
- The integration is pinned to a specific upstream commit and exposes public docx, pptx, and xlsx skills.
- NeMo Gym handles rollout scheduling, while a custom Harbor agent operates on task documents inside Docker.
- A task-local verifier scores every rubric criterion using an OpenAI-compatible judge model, making the setup useful for large-scale agent evaluation.
More from coding & agent
- Repost says Gary Marcus is bullish on neurosymbolic AI, not anti-AI — GaryMarcus · 2026-08-04
- Grok 4.5 Fast gets praise for daily coding, paired with GPT 5.6 Sol for planning — jonathan_wilke · 2026-08-04
- Kaggle and Google Launch $50K Agentic Farm Simulation Competition — WalterReade · 2026-08-04
- NVIDIA open-sources NOOA, a Python object-oriented framework for AI agents — Roger_M_Taylor · 2026-08-04
- AI + Emulator Memory Tool Generates 10,000 Lines of Bit-Identical C Code — yacineMTB · 2026-08-04
- Copper launches a $39 Mac app to capture prompts and ideas across ChatGPT and Cursor — shadcn · 2026-08-04