AI Paper Atlas

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

CodeTaste: Can LLMs Generate Human-Level Code Refactorings?

Alex Thillen, Niels Mündler, Veselin Raychev, Martin Vechev

ICML 2026 · Large language models › Agents, code and math · approach: Software engineering agents · 0 citations (OpenAlex)

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 a plan-then-implement stage (ablation)GPT-5.2 · CodeTaste Open Trackalignment score (%): 7.9 -> 15.1Improved
add a plan-then-implement stage (ablation)Claude Sonnet 4.5 · CodeTaste Open Trackalignment score (%): 6.2 -> 10.7Improved
add a plan-then-implement stage (ablation)GPT-5.1 Codex Mini · CodeTaste Open Trackalignment score (%): 7.2 -> 9.8Improved
add a plan-then-implement stage (ablation)MiniMax M2.7 · CodeTaste Open Trackalignment score (%): 7.0 -> 3.4Hurt
add a plan-then-implement stage (ablation)Qwen3 · CodeTaste Open Trackalignment score (%): 2.0 -> 2.1No clear effect
add oracle selection among multiple plans (ablation)GPT-5.2 · CodeTaste Open Trackalignment score (%): 15.1 -> 20.6Improved
add oracle selection among multiple plans (ablation)Claude Sonnet 4.5 · CodeTaste Open Trackalignment score (%): 10.7 -> 10.1No clear effect
add oracle selection among multiple plans (ablation)GPT-5.1 Codex Mini · CodeTaste Open Trackalignment score (%): 9.8 -> 4.2Hurt
add plan stage (test pass rate) (ablation)Qwen3 · CodeTaste Open TrackPass (%): 58.0 -> 35.0Hurt
add plan stage (test pass rate) (ablation)GPT-5.2 · CodeTaste Open TrackPass (%): 87.0 -> 87.0No clear effect
use randomized six-plan judge selection instead of oracle multiplan (ablation)GPT-5.2 · CodeTaste Open TrackIFR (%): 25.5 -> 26.1No clear effect
give detailed refactoring instructions instead of focus area onlyGPT-5.2 · CodeTastealignment score (%): 7.9 -> 69.6Improved

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