AI Paper Atlas

Research map / Large language models / Agents, code and math

CIRBench: Evaluating Large Language Models as LLVM IR Optimizers

Zi Yang, Haifeng Ding, Fei Liu, Yingying Cheng, Han Cheng, Zhilei Chai, Haojie Zhou

ICML 2026 · Large language models › Agents, code and math · approach: Code generation benchmarks

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add verifier error message and location as a hint (ablation)GPT-5 · CIRBench RepairEquiv@1 (%): 74.0 -> 85.5Improved
add verifier error message and location as a hint (ablation)Claude Sonnet 4.5 · CIRBench RepairEquiv@1 (%): 66.0 -> 76.5Improved
add verifier error message and location as a hint (ablation)Gemini 2.5 Flash · CIRBench RepairEquiv@1 (%): 44.0 -> 74.5Improved
add verifier error message and location as a hint (ablation)Deepseek-V3.2-Exp · CIRBench RepairEquiv@1 (%): 39.0 -> 67.0Improved
add verifier hints on hard and multi-error repairs (ablation)strong models · CIRBench Repair (Medium/Hard, multi-error)Equiv@1: 10–30%Improved
sample 5 candidates instead of 1GPT-5 · CIRBench TransformEquiv (%): 73.3 -> 92.7Improved
sample 5 candidates instead of 1Deepseek-V3.2-Exp · CIRBench TransformEquiv (%): 15.3 -> 40.7Improved
sample 5 candidates instead of 1GPT-5 · CIRBench Refactor-NormalEquiv (%): 22.5 -> 29.5Improved
sample 5 candidates instead of 1GPT-5 · CIRBench Refactor-NormalEM (%): 0.0 -> 0.0No clear effect
run LLM output through -O3 (copilot mode) (ablation)GPT-5 · CIRBench Transform (perf)median dynamic speedup vs -O3: 0.85 -> 0.99Improved
run LLM output through -O3 (copilot mode) (ablation)Grok 4 Fast · CIRBench Transform (perf)median dynamic speedup vs -O3: 0.66 -> 0.98Improved
run LLM output through -O3 (copilot mode) (ablation)GPT-5 · CIRBench Transform (perf)max dynamic speedup vs -O3: 4.89 -> 3.23Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy