Spaces:

---

agent-evals

/

leaderboard

like 0

Build error

Identity: VeigaPunk

We have encrypted GAIA smoke traces ready for community listing:

• PR#4 gpt-4.1-mini 1/1: https://huggingface.co/datasets/agent-evals/hal_traces/discussions/4
• PR#5 xai/grok-4.5 1/1: https://huggingface.co/datasets/agent-evals/hal_traces/discussions/5
• Write access request: https://huggingface.co/datasets/agent-evals/hal_traces/discussions/3

HAL GAIA table remains paused (https://hal.cs.princeton.edu/gaia). harness archived. Can community runs be scored/listed, or when do updates reopen?

Thanks — happy to run full 165-task GAIA under your preferred process.