# ExecVerify: White-Box RL with Verifiable Stepwise Rewards for Code Execution Reasoning

arXiv: https://arxiv.org/abs/2603.11226 · HTML: https://mlsanity.com/paper/2603.11226/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 29 of its 43 references that are in the map.
- First posted on arXiv 11 Mar 2026.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)

## Builds on (most cited first, 29 of 29)

- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code](https://mlsanity.com/paper/2403.07974/) (2024)
- [LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models](https://mlsanity.com/paper/2403.13372/) (2024)
- [Qwen2.5-Coder Technical Report](https://mlsanity.com/paper/2409.12186/) (2024)
- [Teaching Large Language Models to Self-Debug](https://mlsanity.com/paper/2304.05128/) (2023)
- [Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?](https://mlsanity.com/paper/2504.13837/) (2025)
- [Measuring Coding Challenge Competence With APPS](https://mlsanity.com/paper/2105.09938/) (2021)
- [WizardCoder: Empowering Code Large Language Models with Evol-Instruct](https://mlsanity.com/paper/2306.08568/) (2023)
- [A Survey on Large Language Models for Code Generation](https://mlsanity.com/paper/2406.00515/) (2024)
- [BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions](https://mlsanity.com/paper/2406.15877/) (2024)
- [Process Reinforcement through Implicit Rewards](https://mlsanity.com/paper/2502.01456/) (2025)
- [UniXcoder: Unified Cross-Modal Pre-training for Code Representation](https://mlsanity.com/paper/2203.03850/) (2022)
- [The Stack: 3 TB of permissively licensed source code](https://mlsanity.com/paper/2211.15533/) (2022)
- [DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence](https://mlsanity.com/paper/2406.11931/) (2024)
- [CRUXEval: A Benchmark for Code Reasoning, Understanding and Execution](https://mlsanity.com/paper/2401.03065/) (2024)
- [Magicoder: Empowering Code Generation with OSS-Instruct](https://mlsanity.com/paper/2312.02120/) (2023)
- [TACO: Topics in Algorithmic COde generation dataset](https://mlsanity.com/paper/2312.14852/) (2023)
- [REASONING GYM: Reasoning Environments for Reinforcement Learning with Verifiable Rewards](https://mlsanity.com/paper/2505.24760/) (2025)
- [KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Coding](https://mlsanity.com/paper/2503.02951/) (2025)
- [NExT: Teaching Large Language Models to Reason about Code Execution](https://mlsanity.com/paper/2404.14662/) (2024)
- [CodeI/O: Condensing Reasoning Patterns via Code Input-Output Prediction](https://mlsanity.com/paper/2502.07316/) (2025)
- [Reasoning Runtime Behavior of a Program with LLM: How Far Are We?](https://mlsanity.com/paper/2403.16437/) (2024)
- [Two-stage LLM Fine-tuning with Less Specialization and More Generalization](https://mlsanity.com/paper/2211.00635/) (2022)
- [Code Execution with Pre-trained Language Models](https://mlsanity.com/paper/2305.05383/) (2023)
- [CoRe: Benchmarking LLMs Code Reasoning Capabilities through Static Analysis Tasks](https://mlsanity.com/paper/2507.05269/) (2025)
- [Selective Self-to-Supervised Fine-Tuning for Generalization in Large Language Models](https://mlsanity.com/paper/2502.08130/) (2025)
- [Code Execution as Grounded Supervision for LLM Reasoning](https://mlsanity.com/paper/2506.10343/) (2025)
- [Chain of Execution Supervision Promotes General Reasoning in Large Language Models](https://mlsanity.com/paper/2510.23629/) (2025)

## Built on by (most cited first, 6 of 6)

- [Code as Agent Harness](https://mlsanity.com/paper/2605.18747/) (2026)
- [Enhancing the Code Reasoning Capabilities of LLMs via Consistency-based Reinforcement Learning](https://mlsanity.com/paper/2605.17958/) (2026)
- [DiDPO: Diff-in-Diff Policy Optimization for Coding Agent Training](https://mlsanity.com/paper/2608.07147/) (2026)
- [Improving Large Language Models for Code through Runtime Program-State Reasoning](https://mlsanity.com/paper/2609.34359/) (2026)
- [Overlap, Unique and Conflict: Can LLMs Extract What They Can Recognize?](https://mlsanity.com/paper/2609.38799/) (2026)
- [ANTE: Anticipate, Then Execute -- Learning Internal World Models for Coding Agents from Verified Beliefs](https://mlsanity.com/paper/2610.15368/) (2026)

## Abstract (arXiv)

Code LLMs still struggle with code execution reasoning, especially in smaller models. Existing methods rely on supervised fine-tuning (SFT) with teacher-generated explanations, primarily in two forms: (1) input-output (I/O) prediction chains and (2) natural-language descriptions of execution traces. However, intermediate execution steps cannot be explicitly verified during SFT, so the training objective can reduce to merely matching teacher explanations. Moreover, training data is typically collected without explicit control over task difficulty. We introduce ExecVerify, which goes beyond text imitation by incorporating verifiable white-box rewards derived from execution traces, including next-statement prediction and variable value/type prediction. Our work first builds a dataset with multiple difficulty levels via constraint-based program synthesis. Then, we apply reinforcement learning (RL) to reward correct answers about both intermediate execution steps and final outputs, aligning the training objective with semantic correctness at each execution step. Finally, we adopt a two-stage training pipeline that first enhances execution reasoning and then transfers to code generation. Experiments demonstrate that a 7B model trained with ExecVerify achieves performance comparable to 32B models on code reasoning benchmarks and improves pass@1 by up to 5.9% on code generation tasks over strong post-training baselines.