We lifted 7 functions out of this paper's own repositories and ran 6 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| gso-bench/gso | canonical | 6 of 7 |
| Function | Status | Where it lives |
|---|---|---|
| load_gso_predictions | Ran | gso-bench/gso/src/gso/utils/io.py code served (permissive licence) · get_code("3d011bdc974c5e1e") |
| merge_reports | Ran | gso-bench/gso/src/gso/harness/opt_at_k.py code served (permissive licence) · get_code("33002aaa5df9b68d") |
| natural_sort_key | Ran | gso-bench/gso/src/gso/harness/utils.py code served (permissive licence) · get_code("78b4b1c42711889b") |
| prepare_install_commands | Ran | gso-bench/gso/src/gso/collect/utils.py code served (permissive licence) · get_code("8333d7a7f5c976b2") |
| prepare_prob_script | Ran | gso-bench/gso/src/gso/collect/utils.py code served (permissive licence) · get_code("1217b6c996dfb62b") |
| setup_logger | Ran | gso-bench/gso/src/gso/harness/utils.py code served (permissive licence) · get_code("ec3a8a5f64892160") |
| run_evaluation | Not yet run | gso-bench/gso/src/gso/harness/opt_at_k.py code served (permissive licence) · get_code("7d27e938dc504ce5") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Developing high-performance software is a complex task that requires specialized expertise. We introduce GSO, a benchmark for evaluating language models' capabilities in developing high-performance software. We develop an automated pipeline that generates and executes performance tests to analyze repository commit histories to identify 102 challenging optimization tasks across 10 codebases, spanning diverse domains and programming languages. An agent is provided with a codebase and performance test as a precise specification, and tasked to improve the runtime efficiency, which is measured against the expert developer optimization. Our quantitative evaluation reveals that leading SWE-Agents struggle significantly, achieving less than 5% success rate, with limited improvements even with inference-time scaling. Our qualitative analysis identifies key failure modes, including difficulties with low-level languages, practicing lazy optimization strategies, and challenges in accurately localizing bottlenecks. We release the code and artifacts of our benchmark along with agent trajectories to enable future research.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2505.23671")
get_code_for_paper("2505.23671")
have("2505.23671")
Connect an agent — have() is free.