1 min lesson
The prompts in CursorBench are deliberately short and under-specified
Say what this means in practice: "CursorBench hands it a short, vague request."
Step 1 of 2
The prompts in CursorBenchCursor's internal benchmark that scores models on both task performance and token efficiency, not accuracy alone. Press Enter for the full definition. are deliberately short and under-specified. A real developer rarely writes a spec; they type a terse request and expect the agent to figure out what they meant. CursorBench keeps that shape on purpose, so resolving the ambiguity and inferring intent is part of what's being scored, not a detail abstracted away.
Most benchmarks hand the model a fully-specified problem and measure whether it can execute. CursorBenchCursor's internal benchmark that scores models on both task performance and token efficiency, not accuracy alone. Press Enter for the full definition. hands it a short, vague request - the way a real engineer actually asks - and measures whether it can infer intent and resolve the ambiguity before executing. That inference is exactly the non-verifiable skill from section one: no test encodes "did the agent understand what I actually wanted." By keeping prompts under-specified, the benchmark scores the thing that's hardest to score and hardest to game by memorization.
CursorBenchCursor's internal benchmark that scores models on both task performance and token efficiency, not accuracy alone. Press Enter for the full definition. scores two things at once: the quality of the solution and the completion-token length the agent used to get there. Quality alone would let a model ramble its way to a correct answer; pairing it with length makes brevity part of what "good" means, so a solution that's right and tight beats one that's right and bloated.
Learn more
Full explanation
CursorBench vs SWE-Bench
Interactive diagram. Tab through its regions; each focused region shows its detail in the panel below.
Same end goal - measure whether the agent can do real coding work - but CursorBench is built to resist the failure modes that make SWE-bench movements hard to trust.