We present the Comprehensive Verilog Design Problems (CVDP) benchmark, a new dataset and infrastructure to advance LLM and agent research in hardware design and verification. CVDP includes 783 problems across 13 task categories, covering RTL generation, verification, debugging, specification alignment, and technical Q&A authored by experienced hardware engineers. Problems are offered in both non-agentic and agentic formats. The benchmark introduces more realistic and challenging contexts than prior work, with state-of-the-art models achieving no more than 34% pass@1 on code generation. Agentic tasks$\unicode{x2013}$especially those involving RTL reuse and verification$\unicode{x2013}$are particularly difficult. Evaluation uses open-source tools and model scoring infrastructure, with comprehension tasks assessed via BLEU and LLM-based judging. CVDP reveals substantial gaps in current model capabilities, underscoring the need for continued research toward robust, real-world hardware design automation.
LLM-assisted hardware verification is gaining substantial attention due to its potential to significantly reduce the cost and effort of crafting effective testbenches. It also serves as a critical enabler for LLM-aided end-to-end hardware language design. However, existing current LLMs often struggle with Register Transfer Level (RTL) code generation, resulting in testbenches that exhibit functional errors in Hardware Description Languages (HDL) logic. Motivated by the strong performance of LLMs in Python code generation under inference-time sampling strategies, and their promising capabilities as judge agents, we propose PRO-V a fully program generation multi-agent system for robust RTL verification. Pro-V incorporates an efficient best-of-n iterative sampling strategy to enhance the correctness of generated testbenches. Moreover, it introduces an LLM-as-a-judge aid validation framework featuring an automated prompt generation pipeline. By converting rule-based static analysis from the compiler into natural language through in-context learning, this pipeline enables LLMs to assist the compiler in determining whether verification failures stem from errors in the RTL design or the testbench. PRO-V attains a verification accuracy of 87.17% on golden RTL implementations and 76.28% on RTL mutants. Our code is open-sourced at https://github.com/stable-lab/Pro-V.
Large Language Models (LLMs) have made significant strides in front-end code generation. However, existing benchmarks exhibit several critical limitations: many tasks are overly simplistic, test cases often lack rigor, and end-to-end validation is absent. These issues hinder the accurate assessment of model performance. To address these challenges, we present FrontendBench, a benchmark co-developed by humans and LLMs. FrontendBench categorizes tasks based on code functionality and incorporates interactive test scenarios, enabling a more comprehensive and practical evaluation of front-end code generation capabilities. The benchmark comprises 148 meticulously crafted prompt-test case pairs spanning five levels of web components, from basic UI elements to complex interactive features. Each task reflects realistic front-end development challenges. Furthermore, we introduce an automatic evaluation framework that executes generated code within a sandbox environment and assesses outcomes using predefined test scripts. This framework achieves a 90.54% agreement rate with expert human evaluations, demonstrating high reliability. We benchmark several state-of-the-art LLMs on FrontendBench and observe substantial performance disparities in handling real-world front-end tasks. These results highlight FrontendBench as a reliable and scalable benchmark, supporting consistent multimodal evaluation and providing a robust foundation for future research in front-end code generation. Our data and code will be released soon.
The ongoing evolution of language models has led to the development of large-scale architectures that demonstrate exceptional performance across a wide range of tasks. However, these models come with significant computational and energy demands, as well as potential privacy implications. In this context, Small Reasoning Language Models (SRLMs) with approximately 0.5 billion parameters present a compelling alternative due to their remarkable computational efficiency and cost effectiveness, particularly in resource-constrained environments. Despite these advantages, the limited capacity of 0.5 billion parameter models poses challenges in handling complex tasks such as mathematical reasoning and code generation. This research investigates various training strategies, including supervised fine-tuning (SFT), knowledge distillation (KD), and reinforcement learning (RL), as well as their hybrid implementations, to enhance the performance of 0.5B SRLMs. We analyze effective methodologies to bridge the performance gap between SRLMS and larger models and present insights into optimal training pipelines tailored for these smaller architectures. Through extensive experimental validation and analysis, our work aims to provide actionable recommendations for maximizing the reasoning capabilities of 0.5B models.
Enhancing the reasoning capabilities of Large Language Models (LLMs) with efficiency and scalability remains a fundamental challenge in artificial intelligence research. This paper presents a rigorous experimental investigation into how difficulty-aware staged reinforcement learning (RL) strategies can substantially improve LLM reasoning performance. Through systematic analysis, we demonstrate that strategically selecting training data according to well-defined difficulty levels markedly enhances RL optimization. Moreover, we introduce a staged training methodology, progressively exposing models to increasingly challenging tasks, further amplifying reasoning capabilities. Our findings reveal significant cross-domain benefits when simultaneously training models on mathematical reasoning and code generation tasks. Notably, our proposed approach enables a 1.5B parameter model to achieve an accuracy of 42.3\% on the AIME-2024 benchmark, 89.5\% on the MATH-500 benchmark. These results underscore the efficacy of our method in advancing the reasoning proficiency of LLMs. We will open-source our datasets on GitHub and Hugging Face.
Finetuning is a critical step for adapting large language models (LLMs) to domain-specific downstream tasks. To mitigate the substantial computational and memory costs of full-model fine-tuning, Parameter-Efficient Finetuning (PEFT) methods have been proposed to update only a small subset of model parameters. However, performance gaps between PEFT approaches and full-model fine-tuning still exist. In this work, we present DiaBlo, a simple yet effective PEFT approach that updates only the diagonal blocks of selected model weight matrices. Unlike Low Rank Adaptation (LoRA) and its variants, DiaBlo eliminates the need for low rank matrix products, thereby avoiding the reliance on auxiliary initialization schemes or customized optimization strategies to improve convergence. This design leads to stable and robust convergence while maintaining comparable memory efficiency and training speed to LoRA. We conduct extensive experiments across a range of tasks, including commonsense reasoning, arithmetic reasoning, code generation, and safety alignment, to evaluate the effectiveness and efficiency of DiaBlo. Across these benchmarks, DiaBlo demonstrates strong and consistent performance while maintaining high memory efficiency and fast finetuning speed. Codes are available at https://github.com/ziyangjoy/DiaBlo.
Large Language Models (LLMs) are rapidly advancing across diverse domains, yet their application in theoretical physics research is not yet mature. This position paper argues that LLM agents can potentially help accelerate theoretical, computational, and applied physics when properly integrated with domain knowledge and toolbox. We analyze current LLM capabilities for physics -- from mathematical reasoning to code generation -- identifying critical gaps in physical intuition, constraint satisfaction, and reliable reasoning. We envision future physics-specialized LLMs that could handle multimodal data, propose testable hypotheses, and design experiments. Realizing this vision requires addressing fundamental challenges: ensuring physical consistency, and developing robust verification methods. We call for collaborative efforts between physics and AI communities to help advance scientific discovery in physics.
This position paper proposes a fundamental shift in designing code generation models: treating reasoning depth as a controllable resource. Rather than being an incidental byproduct of prompting, we argue that the trade-off between rapid, direct answers ("fast thinking") and elaborate, chain-of-thought deliberation ("slow thinking") must be explicitly managed. We contend that optimizing reasoning budgets across the entire model lifecycle - from synthetic data creation and benchmarking to real-world deploymen - can unlock superior trade-offs among accuracy, latency, and cost. This paper outlines how adaptive control over reasoning can enrich supervision signals, motivate new multi-dimensional benchmarks, and inform cost-aware, security-conscious deployment policies. By viewing fast and slow thinking as complementary modes to be scheduled, we envision coding agents that think deep when necessary and act fast when possible.
The advent of Large Language Models (LLMs) has spurred the development of coding agents for real-world code generation. As a widely used benchmark for evaluating the code generation capabilities of these agents, SWE-Bench uses real-world problems based on GitHub issues and their corresponding pull requests. However, the manually written test cases included in these pull requests are often insufficient, allowing generated patches to pass the tests without resolving the underlying issue. To address this challenge, we introduce UTGenerator, an LLM-driven test case generator that automatically analyzes codebases and dependencies to generate test cases for real-world Python projects. Building on UTGenerator, we propose UTBoost, a comprehensive framework for test case augmentation. In our evaluation, we identified 36 task instances with insufficient test cases and uncovered 345 erroneous patches incorrectly labeled as passed in the original SWE Bench. These corrections, impacting 40.9% of SWE-Bench Lite and 24.4% of SWE-Bench Verified leaderboard entries, yield 18 and 11 ranking changes, respectively.
Large Language Models (LLMs) have demonstrated remarkable capabilities in code generation, capable of tackling complex tasks during inference. However, the extent to which LLMs can be utilized for code checking or debugging through test case generation remains largely unexplored. We investigate this problem from the perspective of competition-level programming (CP) programs and propose TCGBench, a Benchmark for (LLM generation of) Test Case Generators. This benchmark comprises two tasks, aimed at studying the capabilities of LLMs in (1) generating valid test case generators for a given CP problem, and further (2) generating targeted test case generators that expose bugs in human-written code. Experimental results indicate that while state-of-the-art LLMs can generate valid test case generators in most cases, most LLMs struggle to generate targeted test cases that reveal flaws in human code effectively. Especially, even advanced reasoning models (e.g., o3-mini) fall significantly short of human performance in the task of generating targeted generators. Furthermore, we construct a high-quality, manually curated dataset of instructions for generating targeted generators. Analysis demonstrates that the performance of LLMs can be enhanced with the aid of this dataset, by both prompting and fine-tuning.
Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities in automated front-end engineering, e.g., generating UI code from visual designs. However, existing front-end UI code generation benchmarks have the following limitations: (1) While framework-based development becomes predominant in modern front-end programming, current benchmarks fail to incorporate mainstream development frameworks. (2) Existing evaluations focus solely on the UI code generation task, whereas practical UI development involves several iterations, including refining editing, and repairing issues. (3) Current benchmarks employ unidimensional evaluation, lacking investigation into influencing factors like task difficulty, input context variations, and in-depth code-level analysis. To bridge these gaps, we introduce DesignBench, a multi-framework, multi-task evaluation benchmark for assessing MLLMs' capabilities in automated front-end engineering. DesignBench encompasses three widely-used UI frameworks (React, Vue, and Angular) alongside vanilla HTML/CSS, and evaluates on three essential front-end tasks (generation, edit, and repair) in real-world development workflows. DesignBench contains 900 webpage samples spanning over 11 topics, 9 edit types, and 6 issue categories, enabling detailed analysis of MLLM performance across multiple dimensions. Our systematic evaluation reveals critical insights into MLLMs' framework-specific limitations, task-related bottlenecks, and performance variations under different conditions, providing guidance for future research in automated front-end development. Our code and data are available at https://github.com/WebPAI/DesignBench.
Infrastructure-as-Code (IaC) generation holds significant promise for automating cloud infrastructure provisioning. Recent advances in Large Language Models (LLMs) present a promising opportunity to democratize IaC development by generating deployable infrastructure templates from natural language descriptions, but current evaluation focuses on syntactic correctness while ignoring deployability, the fatal measure of IaC template utility. We address this gap through two contributions: (1) IaCGen, an LLM-based deployability-centric framework that uses iterative feedback mechanism to generate IaC templates, and (2) DPIaC-Eval, a deployability-centric IaC template benchmark consists of 153 real-world scenarios that can evaluate syntax, deployment, user intent, and security. Our evaluation reveals that state-of-the-art LLMs initially performed poorly, with Claude-3.5 and Claude-3.7 achieving only 30.2% and 26.8% deployment success on the first attempt respectively. However, IaCGen transforms this performance dramatically: all evaluated models reach over 90% passItr@25, with Claude-3.5 and Claude-3.7 achieving 98% success rate. Despite these improvements, critical challenges remain in user intent alignment (25.2% accuracy) and security compliance (8.4% pass rate), highlighting areas requiring continued research. Our work provides the first comprehensive assessment of deployability-centric IaC template generation and establishes a foundation for future research.
Constructing real-world data-to-insight pipelines often involves data extraction from data lakes, data integration across heterogeneous data sources, and diverse operations from data cleaning to analysis. The design and implementation of data science pipelines require domain knowledge, technical expertise, and even project-specific insights. AI systems have shown remarkable reasoning, coding, and understanding capabilities. However, it remains unclear to what extent these capabilities translate into successful design and execution of such complex pipelines. We introduce KRAMABENCH: a benchmark composed of 104 manually-curated real-world data science pipelines spanning 1700 data files from 24 data sources in 6 different domains. We show that these pipelines test the end-to-end capabilities of AI systems on data processing, requiring data discovery, wrangling and cleaning, efficient processing, statistical reasoning, and orchestrating data processing steps given a high-level task. Our evaluation tests 5 general models and 3 code generation models using our reference framework, DS-GURU, which instructs the AI model to decompose a question into a sequence of subtasks, reason through each step, and synthesize Python code that implements the proposed design. Our results on KRAMABENCH show that, although the models are sufficiently capable of solving well-specified data science code generation tasks, when extensive data processing and domain knowledge are required to construct real-world data science pipelines, existing out-of-box models fall short. Progress on KramaBench represents crucial steps towards developing autonomous data science agents for real-world applications. Our code, reference framework, and data are available at https://github.com/mitdbg/KramaBench.
The code generation capabilities of large language models(LLMs) have emerged as a critical dimension in evaluating their overall performance. However, prior research has largely overlooked the security risks inherent in the generated code. In this work, we introduce SafeGenBench, a benchmark specifically designed to assess the security of LLM-generated code. The dataset encompasses a wide range of common software development scenarios and vulnerability types. Building upon this benchmark, we develop an automatic evaluation framework that leverages both static application security testing(SAST) and LLM-based judging to assess the presence of security vulnerabilities in model-generated code. Through the empirical evaluation of state-of-the-art LLMs on SafeGenBench, we reveal notable deficiencies in their ability to produce vulnerability-free code. Our findings highlight pressing challenges and offer actionable insights for future advancements in the secure code generation performance of LLMs. The data and code will be released soon.
Recent advancements in LLMs indicate potential for novel applications, e.g., through reasoning capabilities in the latest OpenAI and DeepSeek models. For applying these models in specific domains beyond text generation, LLM-based multi-agent approaches can be utilized that solve complex tasks by combining reasoning techniques, code generation, and software execution. Applications might utilize these capabilities and the knowledge of specialized LLM agents. However, while many evaluations are performed on LLMs, reasoning techniques, and applications individually, their joint specification and combined application is not explored well. Defined specifications for multi-agent LLM systems are required to explore their potential and their suitability for specific applications, allowing for systematic evaluations of LLMs, reasoning techniques, and related aspects. This paper reports the results of exploratory research to specify and evaluate these aspects through a multi-agent system. The system architecture and prototype are extended from previous research and a specification is introduced for multi-agent systems. Test cases involving cybersecurity tasks indicate feasibility of the architecture and evaluation approach. In particular, the results show the evaluation of question answering, server security, and network security tasks that were completed correctly by agents with LLMs from OpenAI and DeepSeek.
The widespread adoption of Large Language Models (LLMs) has heightened concerns about their security, particularly their vulnerability to jailbreak attacks that leverage crafted prompts to generate malicious outputs. While prior research has been conducted on general security capabilities of LLMs, their specific susceptibility to jailbreak attacks in code generation remains largely unexplored. To fill this gap, we propose MalwareBench, a benchmark dataset containing 3,520 jailbreaking prompts for malicious code-generation, designed to evaluate LLM robustness against such threats. MalwareBench is based on 320 manually crafted malicious code generation requirements, covering 11 jailbreak methods and 29 code functionality categories. Experiments show that mainstream LLMs exhibit limited ability to reject malicious code-generation requirements, and the combination of multiple jailbreak methods further reduces the model's security capabilities: specifically, the average rejection rate for malicious content is 60.93%, dropping to 39.92% when combined with jailbreak attack algorithms. Our work highlights that the code security capabilities of LLMs still pose significant challenges.
We present a novel approach to neural code generation that incorporates real-time execution signals into the language model generation process. While large language models (LLMs) have demonstrated impressive code generation capabilities, they typically do not utilize execution feedback during inference, a critical signal that human programmers regularly leverage. Our method, Execution-Guided Classifier-Free Guidance (EG-CFG), dynamically incorporates execution signals as the model generates code, providing line-by-line feedback that guides the generation process toward executable solutions. EG-CFG employs a multi-stage process: first, we conduct beam search to sample candidate program completions for each line; second, we extract execution signals by executing these candidates against test cases; and finally, we incorporate these signals into the prompt during generation. By maintaining consistent signals across tokens within the same line and refreshing signals at line boundaries, our approach provides coherent guidance while preserving syntactic structure. Moreover, the method naturally supports native parallelism at the task level in which multiple agents operate in parallel, exploring diverse reasoning paths and collectively generating a broad set of candidate solutions. Our experiments across diverse coding tasks demonstrate that EG-CFG significantly improves code generation performance compared to standard approaches, achieving state-of-the-art results across various levels of complexity, from foundational problems to challenging competitive programming tasks. Our code is available at: https://github.com/boazlavon/eg_cfg
Code generation is one of the most active areas of application of Large Language Models (LLMs). While LLMs lower barriers to writing code and accelerate development process, the overall quality of generated programs depends on the quality of given prompts. Specifically, functionality and quality of generated code can be sensitive to user's background and familiarity with software development. It is therefore important to quantify LLM's sensitivity to variations in the input. To this end we propose a synthetic evaluation pipeline for code generation with LLMs, as well as a systematic persona-based evaluation approach to expose qualitative differences of LLM responses dependent on prospective user background. Both proposed methods are completely independent from specific programming tasks and LLMs, and thus are widely applicable. We provide experimental evidence illustrating utility of our methods and share our code for the benefit of the community.
Reinforcement learning (RL) has shown great effectiveness for fine-tuning large language models (LLMs) using tasks that are challenging yet easily verifiable, such as math reasoning or code generation. However, extending this success to visual perception in vision-language models (VLMs) has been impeded by the scarcity of vision-centric tasks that are simultaneously challenging and unambiguously verifiable. To this end, we introduce ViCrit (Visual Caption Hallucination Critic), an RL proxy task that trains VLMs to localize a subtle, synthetic visual hallucination injected into paragraphs of human-written image captions. Starting from a 200-word captions, we inject a single, subtle visual description error-altering a few words on objects, attributes, counts, or spatial relations-and task the model to pinpoint the corrupted span given the image and the modified caption. This formulation preserves the full perceptual difficulty while providing a binary, exact-match reward that is easy to compute and unambiguous. Models trained with the ViCrit Task exhibit substantial gains across a variety of VL benchmarks. Crucially, the improvements transfer beyond natural-image training data to abstract image reasoning and visual math, showing promises of learning to perceive rather than barely memorizing seen objects. To facilitate evaluation, we further introduce ViCrit-Bench, a category-balanced diagnostic benchmark that systematically probes perception errors across diverse image domains and error types. Together, our results demonstrate that fine-grained hallucination criticism is an effective and generalizable objective for enhancing visual perception in VLMs.
Large Language Models (LLMs) have achieved remarkable success in tasks requiring complex reasoning, such as code generation, mathematical problem solving, and algorithmic synthesis -- especially when aided by reasoning tokens and Chain-of-Thought prompting. Yet, a core question remains: do these models truly reason, or do they merely exploit shallow statistical patterns? In this paper, we introduce Chain-of-Code Collapse, where we systematically investigate the robustness of reasoning LLMs by introducing a suite of semantically faithful yet adversarially structured prompt perturbations. Our evaluation -- spanning 700 perturbed code generations derived from LeetCode-style problems -- applies transformations such as storytelling reframing, irrelevant constraint injection, example reordering, and numeric perturbation. We observe that while certain modifications severely degrade performance (with accuracy drops up to -42.1%), others surprisingly improve model accuracy by up to 35.3%, suggesting sensitivity not only to semantics but also to surface-level prompt dynamics. These findings expose the fragility and unpredictability of current reasoning systems, underscoring the need for more principles approaches to reasoning alignments and prompting robustness. We release our perturbation datasets and evaluation framework to promote further research in trustworthy and resilient LLM reasoning.