# Kevin: Multi-Turn RL for Generating CUDA Kernels

arXiv: https://arxiv.org/abs/2507.11948 · HTML: https://mlsanity.com/paper/2507.11948/

## Key facts

- Cited by 78 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 43 references that are in the map.
- First posted on arXiv 16 Jul 2025.

## Builds on (most cited first, 20 of 20)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [Concrete Problems in AI Safety](https://mlsanity.com/paper/1606.06565/) (2016)
- [Large Language Monkeys: Scaling Inference Compute with Repeated Sampling](https://mlsanity.com/paper/2407.21787/) (2024)
- [Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?](https://mlsanity.com/paper/2504.13837/) (2025)
- [Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm](https://mlsanity.com/paper/1712.01815/) (2017)
- [RL2: Fast Reinforcement Learning via Slow Reinforcement Learning](https://mlsanity.com/paper/1611.02779/) (2016)
- [Reinforcement Learning for Reasoning in Large Language Models with One Training Example](https://mlsanity.com/paper/2504.20571/) (2025)
- [FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving](https://mlsanity.com/paper/2501.01005/) (2025)
- [Optimizing Test-Time Compute via Meta Reinforcement Fine-Tuning](https://mlsanity.com/paper/2503.07572/) (2025)
- [TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators](https://mlsanity.com/paper/2502.14752/) (2025)
- [ZeRO++: Extremely Efficient Collective Communication for Giant Model Training](https://mlsanity.com/paper/2306.10209/) (2023)
- [LLMs are Greedy Agents: Effects of RL Fine-tuning on Decision-Making Abilities](https://mlsanity.com/paper/2504.16078/) (2025)
- [Putting the Value Back in RL: Better Test-Time Scaling by Unifying LLM Reasoners With Verifiers](https://mlsanity.com/paper/2505.04842/) (2025)

## Built on by (most cited first, 30 of 78)

- [Astra: A Multi-Agent System for GPU Kernel Performance Optimization](https://mlsanity.com/paper/2509.07506/) (2025)
- [CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation](https://mlsanity.com/paper/2602.24286/) (2026)
- [CudaForge: An Agent Framework with Hardware Feedback for CUDA Kernel Optimization](https://mlsanity.com/paper/2511.01884/) (2025)
- [MultiKernelBench: A Multi-Platform Benchmark for Kernel Generation](https://mlsanity.com/paper/2507.17773/) (2025)
- [FlashInfer-Bench: Building the Virtuous Cycle for AI-driven LLM Systems](https://mlsanity.com/paper/2601.00227/) (2026)
- [TritonRL: Training LLMs to Think and Code Triton Without Cheating](https://mlsanity.com/paper/2510.17891/) (2025)
- [K-Search: LLM Kernel Generation via Co-Evolving Intrinsic World Model](https://mlsanity.com/paper/2602.19128/) (2026)
- [AccelOpt: A Self-Improving LLM Agentic System for AI Accelerator Kernel Optimization](https://mlsanity.com/paper/2511.15915/) (2025)
- [Dr. Kernel: Reinforcement Learning Done Right for Triton Kernel Generations](https://mlsanity.com/paper/2602.05885/) (2026)
- [ConCuR: Conciseness Makes State-of-the-Art Kernel Generation](https://mlsanity.com/paper/2510.07356/) (2025)
- [CUDA-L2: Surpassing cuBLAS Performance for Matrix Multiplication through Reinforcement Learning](https://mlsanity.com/paper/2512.02551/) (2025)
- [KernelFoundry: Hardware-aware evolutionary GPU kernel optimization](https://mlsanity.com/paper/2603.12440/) (2026)
- [SwizzlePerf: Hardware-Aware LLMs for GPU Kernel Performance Optimization](https://mlsanity.com/paper/2508.20258/) (2025)
- [Scaling Generalist Data-Analytic Agents](https://mlsanity.com/paper/2509.25084/) (2025)
- [Towards Automated Kernel Generation in the Era of LLMs](https://mlsanity.com/paper/2601.15727/) (2026)
- [KernelSkill: A Multi-Agent Framework for GPU Kernel Optimization](https://mlsanity.com/paper/2603.10085/) (2026)
- [From Large to Small: Transferring CUDA Optimization Expertise via Reasoning Graph](https://mlsanity.com/paper/2510.19873/) (2025)
- [Hardening Agent Benchmarks with Adversarial Hacker-Fixer Loops](https://mlsanity.com/paper/2606.08960/) (2026)
- [Understanding the Challenges in Iterative Generative Optimization with LLMs](https://mlsanity.com/paper/2603.23994/) (2026)
- [cuPilot: A Strategy-Coordinated Multi-agent Framework for CUDA Kernel Evolution](https://mlsanity.com/paper/2512.16465/) (2025)
- [AKG kernel Agent: A Multi-Agent Framework for Cross-Platform Kernel Synthesis](https://mlsanity.com/paper/2512.23424/) (2025)
- [AscendCraft: Automatic Ascend NPU Kernel Generation via DSL-Guided Transcompilation](https://mlsanity.com/paper/2601.22760/) (2026)
- [StitchCUDA: An Automated Multi-Agents End-to-End GPU Programing Framework with Rubric-based Agentic Reinforcement Learning](https://mlsanity.com/paper/2603.02637/) (2026)
- [Kernel-Smith: A Unified Recipe for Evolutionary Kernel Optimization](https://mlsanity.com/paper/2603.28342/) (2026)
- [LANPO: Bootstrapping Language and Numerical Feedback for Reinforcement Learning in LLMs](https://mlsanity.com/paper/2510.16552/) (2025)
- [KernelEvolve: Scaling Agentic Kernel Coding for Heterogeneous AI Accelerators at Meta](https://mlsanity.com/paper/2512.23236/) (2025)
- [Improving Efficiency of GPU Kernel Optimization Agents using a Domain-Specific Language and Speed-of-Light Guidance](https://mlsanity.com/paper/2603.29010/) (2026)
- [Optimizing PyTorch Inference with LLM-Based Multi-Agent Systems](https://mlsanity.com/paper/2511.16964/) (2025)
- [MaxCode: A Max-Reward Reinforcement Learning Framework for Automated Code Optimization](https://mlsanity.com/paper/2601.05475/) (2026)
- [Fine-Tuning GPT-5 for GPU Kernel Generation](https://mlsanity.com/paper/2602.11000/) (2026)

## Abstract (arXiv)

Writing GPU kernels is a challenging task and critical for AI systems' efficiency. It is also highly iterative: domain experts write code and improve performance through execution feedback. Moreover, it presents verifiable rewards like correctness and speedup, making it a natural environment to apply Reinforcement Learning (RL). To explicitly incorporate the iterative nature of this process into training, we develop a flexible multi-turn RL recipe that addresses unique challenges encountered in real-world settings, such as learning from long trajectories and effective reward attribution across turns. We present Kevin - K(ernel D)evin, the first model trained with multi-turn RL for CUDA kernel generation and optimization. In our evaluation setup, Kevin shows significant gains over its base model (QwQ-32B), improving correctness of generated kernels (in pure CUDA) from 56% to 82% and mean speedup from 0.53x to 1.10x of baseline (PyTorch Eager), and surpassing frontier models like o4-mini (0.78x). Finally, we study its behavior across test-time scaling axes: we found scaling serial refinement more beneficial than parallel sampling. In particular, when given more refinement turns, Kevin shows a higher rate of improvement.