Unverified50% confidenceFactExact time
Terminal-Bench scores are heavily influenced by the agent scaffold and retry strategy, and cross-model comparisons require confirming that the evaluation conditions are consistent.
1
Sources
50%
Confidence
Long-term
Relevance
10/23/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe evaluation scope of TrialPD-Bench is not limited to final answer correctness, but also includes the evidence recall rate of agent trajectories, agent behavior patterns, and cost.76% similarUnverifiedTerminal Bench tests what a model can do rather than what it knows, measuring the model's complete closed-loop capability of planning, executing, observing feedback, and correcting errors74% similarUnverifiedThe evaluation metrics for LOADBench's inverse sizing design task typically include the final performance compliance rate, the simulation verification pass rate, and the closeness to the true Pareto front.74% similarUnverifiedThis benchmark is not a SWE-bench- or Terminal-Bench-style task but long-cycle agentic building of real applications, and the scoring checks were written by the team itself and may have blind spots.69% similarUnverifiedScopeBench employs a dual-arm scoring mechanism: trajectories under the Scoped condition are first processed by a deterministic validator, and those that fail are then assessed by an Agentic Judge to estimate whether an out-of-scope call occurred.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1091982API
curl https://kongchang.com/api/v1/knowledge/claims/1091982MCP
get_claim(id=1091982)