SWE-Atlas
A programming capability benchmark targeting real-world software engineering tasks, focusing on end-to-end ability to locate and fix issues within actual GitHub repositories. It serves as a key metric for assessing the practical usability of coding agents and represents an evolution of the SWE-bench benchmark.
Timeline (last 90 days)
The vendor officially claims that Weave Router 2.0 performs on par with GPT-6 Astra on the Terminal-Bench 4.0 and SWE-Atlas benchmarks, at half the cost and twice the speed.
Terminal-Bench focuses on the ability to complete multi-step command-line tasks in a terminal environment, while SWE-Atlas (and its predecessor SWE-bench) focuses on the end-to-end ability to locate and fix issues in real GitHub repositories.
All Facts (2)
The vendor officially claims that Weave Router 2.0 performs on par with GPT-6 Astra on the Terminal-Bench 4.0 and SWE-Atlas benchmarks, at half the cost and twice the speed.
50%UnverifiedTerminal-Bench focuses on the ability to complete multi-step command-line tasks in a terminal environment, while SWE-Atlas (and its predecessor SWE-bench) focuses on the end-to-end ability to locate and fix issues in real GitHub repositories.
50%