Unverified50% confidenceFactExact time
Almost all frontier labs track performance on the TerminalBench benchmark.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/20/2026
First Seen
Valid until: 2/18/2027
Sources
Related Entities
Related Claims
UnverifiedThe original Terminal-Bench focuses on command-line operation evaluation.71% similarVerifiedTerminal-Bench is a benchmark specifically designed to evaluate AI models' ability to execute complex tasks in command-line environments, covering scenarios such as file operations, script writing, and system debugging.65% similarUnverifiedTerminal Bench Science 0.1 is a terminal-operation benchmark aimed at scientific research scenarios, requiring models to complete multi-step tasks related to scientific computing in a simulated Linux terminal environment.65% similarUnverifiedTerminal-Bench scores are heavily influenced by the agent scaffold and retry strategy, and cross-model comparisons require confirming that the evaluation conditions are consistent.63% similarUnverifiedTerminal-Bench-Science anchors its test tasks to specific scientific research workflows, requiring agents to independently complete the entire process from problem understanding to result output in a terminal environment containing real tools and data.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1190450API
curl https://kongchang.com/api/v1/knowledge/claims/1190450MCP
get_claim(id=1190450)