CodeARC: Benchmarking Reasoning Capabilities of LLM Agents for Inductive Program Synthesis

• URL: http://arxiv.org/abs/2503.23145v2
• Date: Fri, 08 Aug 2025 07:13:11 GMT
• Title: CodeARC: Benchmarking Reasoning Capabilities of LLM Agents for Inductive Program Synthesis
• Authors: Anjiang Wei, Tarun Suresh, Jiannan Cao, Naveen Kannan, Yuheng Wu, Kai Yan, Thiago S. F. X. Teixeira, Ke Wang, Alex Aiken,
• Abstract summary: Existing evaluation protocols rely on static sets of examples and held-out tests, offering no feedback when synthesized functions are incorrect.<n>We propose CodeARC, a new evaluation framework where agents interact with a hidden target function by querying it with new inputs.<n>We construct the first large-scale benchmark for general-purpose inductive program synthesis, featuring 1114 functions.
• Score: 6.8081984950459
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Inductive program synthesis, or programming by example, requires synthesizing functions from input-output examples that generalize to unseen inputs. While large language model agents have shown promise in programming tasks guided by natural language, their ability to perform inductive program synthesis is underexplored. Existing evaluation protocols rely on static sets of examples and held-out tests, offering no feedback when synthesized functions are incorrect and failing to reflect real-world scenarios such as reverse engineering. We propose CodeARC, the Code Abstraction and Reasoning Challenge, a new evaluation framework where agents interact with a hidden target function by querying it with new inputs, synthesizing candidate functions, and iteratively refining their solutions using a differential testing oracle. This interactive setting encourages agents to perform function calls and self-correction based on feedback. We construct the first large-scale benchmark for general-purpose inductive program synthesis, featuring 1114 functions. Among 18 models evaluated, o3-mini performs best with a success rate of 52.7%, highlighting the difficulty of this task. Fine-tuning LLaMA-3.1-8B-Instruct on curated synthesis traces yields up to a 31% relative performance gain. CodeARC provides a more realistic and challenging testbed for evaluating LLM-based program synthesis and inductive reasoning. Our code, data, and models are publicly available at https://github.com/Anjiang-Wei/CodeARC

Related papers

• RAVEL: Reasoning Agents for Validating and Evaluating LLM Text Synthesis [78.32151470154422]
  We introduce RAVEL, an agentic framework that enables the testers to autonomously plan and execute typical synthesis operations.<n>We present C3EBench, a benchmark comprising 1,258 samples derived from professional human writings.<n>By augmenting RAVEL with SOTA LLMs as operators, we find that such agentic text synthesis is dominated by the LLM's reasoning capability.
  arXiv  Detail & Related papers  (2026-02-28T14:47:34Z)
• SPARC: Scenario Planning and Reasoning for Automated C Unit Test Generation [1.0010193170880752]
  We introduce a neuro-symbolic, scenario-based framework that bridges the gap between high-level program intent and the rigid syntactic constraints of pointer arithmetic and manual memory management.<n>We evaluate it on 59 real-world and algorithmic subjects, where it outperforms the vanilla prompt generation baseline by 31.36% in line coverage, 26.01% in branch coverage, and 20.78% in mutation score, matching or exceeding the symbolic execution tool KLEE.
  arXiv  Detail & Related papers  (2026-02-18T18:09:03Z)
• From Brute Force to Semantic Insight: Performance-Guided Data Transformation Design with LLMs [48.83701310501069]
  Large language models (LLMs) have achieved notable performance in code synthesis.<n>We introduce a performance-aware, closed-loop solution that enables LLMs to autonomously engineer optimal transformations.<n>We fine-tune LLMs with Low-Rank Adaptation on a novel repository of more than 6,000 empirically evaluated PyTorch augmentation functions.
  arXiv  Detail & Related papers  (2026-01-07T11:13:02Z)
• Program Synthesis via Test-Time Transduction [26.30808249424997]
  We introduce transductive program synthesis, a new formulation of the program synthesis task that explicitly leverages test inputs during synthesis.<n>We evaluate our approach on four benchmarks: Playgol, MBPP+, 1D-ARC, and programmatic world modeling on MiniGrid.<n>We demonstrate that our method significantly improves program synthesis in both accuracy and efficiency.
  arXiv  Detail & Related papers  (2025-09-22T06:53:32Z)
• STEPWISE-CODEX-Bench: Evaluating Complex Multi-Function Comprehension and Fine-Grained Execution Reasoning [6.282781900938977]
  We present STEPWISE-CODEX-Bench (SX-Bench), a novel benchmark for complex multi-function understanding and fine-grained execution reasoning.<n>SX-Bench is highly discriminative, even the state-of-the-art OpenAI-O3 achieves only 78.37 percent accuracy on Hard-Reasoning tasks.
  arXiv  Detail & Related papers  (2025-08-07T09:28:43Z)
• Grammar-Guided Evolutionary Search for Discrete Prompt Optimisation [63.97051732013936]
  We propose an evolutionary search approach to automated discrete prompt optimisation consisting of two phases.<n>In the first phase, grammar-guided genetic programming is invoked to synthesise prompt-creating programmes.<n>In the second phase, local search is applied to explore the neighbourhoods of best-performing programmes.
  arXiv  Detail & Related papers  (2025-07-14T14:34:15Z)
• In-context Demonstration Matters: On Prompt Optimization for Pseudo-Supervision Refinement [71.60563181678323]
  Large language models (LLMs) have achieved great success across diverse tasks, and fine-tuning is sometimes needed to further enhance generation quality.<n>To handle these challenges, a direct solution is to generate high-confidence'' data from unsupervised downstream tasks.<n>We propose a novel approach, pseudo-supervised demonstrations aligned prompt optimization (PAPO) algorithm, which jointly refines both the prompt and the overall pseudo-supervision.
  arXiv  Detail & Related papers  (2024-10-04T03:39:28Z)
• Case2Code: Scalable Synthetic Data for Code Generation [105.89741089673575]
  Large Language Models (LLMs) have shown outstanding breakthroughs in code generation.<n>Recent work improves code LLMs by training on synthetic data generated by some powerful LLMs.<n>We propose a textbfCase2Code task by exploiting the expressiveness and correctness of programs.
  arXiv  Detail & Related papers  (2024-07-17T11:35:00Z)
• Inductive Predicate Synthesis Modulo Programs (Extended) [1.7372615815088566]
  A growing trend in program analysis is to encode verification conditions within the language of the input program. We propose Inductive Predicate Synthesis Modulo Programs (IPS-MP) which extends high-level languages with minimal synthesis features.
  arXiv  Detail & Related papers  (2024-07-11T12:51:08Z)
• HYSYNTH: Context-Free LLM Approximation for Guiding Program Synthesis [25.260063704712458]
  Large language models (LLMs) often fail to produce fully correct programs in unfamiliar DSLs. Motivated by these limitations, we introduce a hybrid approach, where LLM completions for a given task are used to learn a task-specific, context-free surrogate model. We evaluate this hybrid approach on three domains, and show that it outperforms both unguided search and direct sampling from LLMs, as well as existing program synthesizers.
  arXiv  Detail & Related papers  (2024-05-24T18:45:51Z)
• Guiding Enumerative Program Synthesis with Large Language Models [15.500250058226474]
  In this paper, we evaluate the abilities of Large Language Models to solve formal synthesis benchmarks. When one-shot synthesis fails, we propose a novel enumerative synthesis algorithm. We find that GPT-3.5 as a stand-alone tool for formal synthesis is easily outperformed by state-of-the-art formal synthesis algorithms.
  arXiv  Detail & Related papers  (2024-03-06T19:13:53Z)
• LINC: A Neurosymbolic Approach for Logical Reasoning by Combining Language Models with First-Order Logic Provers [60.009969929857704]
  Logical reasoning is an important task for artificial intelligence with potential impacts on science, mathematics, and society. In this work, we reformulating such tasks as modular neurosymbolic programming, which we call LINC. We observe significant performance gains on FOLIO and a balanced subset of ProofWriter for three different models in nearly all experimental conditions we evaluate.
  arXiv  Detail & Related papers  (2023-10-23T17:58:40Z)
• ExeDec: Execution Decomposition for Compositional Generalization in Neural Program Synthesis [54.18659323181771]
  We characterize several different forms of compositional generalization that are desirable in program synthesis. We propose ExeDec, a novel decomposition-based strategy that predicts execution subgoals to solve problems step-by-step informed by program execution at each step.
  arXiv  Detail & Related papers  (2023-07-26T01:07:52Z)
• Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control [23.115574119132507]
  Building agents with large language models for computer control is a burgeoning research area, where the agent receives computer states and performs actions to complete tasks. Previous computer agents have demonstrated the benefits of in-context learning (ICL), but their performance is hindered by several issues. We introduce Synapse, a computer agent featuring three key components: i) state abstraction, which filters out task-irrelevant information from raw states, allowing more exemplars within the limited context, ii) trajectory-as-exemplar prompting, which prompts the LLM with complete trajectories of the abstracted states and actions to improve multi-step decision-
  arXiv  Detail & Related papers  (2023-06-13T15:49:41Z)
• Latent Execution for Neural Program Synthesis Beyond Domain-Specific Languages [97.58968222942173]
  We take the first step to synthesize C programs from input-output examples. In particular, we propose La Synth, which learns the latent representation to approximate the execution of partially generated programs. We show that training on these synthesized programs further improves the prediction performance for both Karel and C program synthesis.
  arXiv  Detail & Related papers  (2021-06-29T02:21:32Z)
• Leveraging Language to Learn Program Abstractions and Search Heuristics [66.28391181268645]
  We introduce LAPS (Language for Abstraction and Program Search), a technique for using natural language annotations to guide joint learning of libraries and neurally-guided search models for synthesis. When integrated into a state-of-the-art library learning system (DreamCoder), LAPS produces higher-quality libraries and improves search efficiency and generalization.
  arXiv  Detail & Related papers  (2021-06-18T15:08:47Z)
• BUSTLE: Bottom-Up Program Synthesis Through Learning-Guided Exploration [72.88493072196094]
  We present a new synthesis approach that leverages learning to guide a bottom-up search over programs. In particular, we train a model to prioritize compositions of intermediate values during search conditioned on a set of input-output examples. We show that the combination of learning and bottom-up search is remarkably effective, even with simple supervised learning approaches.
  arXiv  Detail & Related papers  (2020-07-28T17:46:18Z)
• Synthesize, Execute and Debug: Learning to Repair for Neural Program Synthesis [81.54148730967394]
  We propose SED, a neural program generation framework that incorporates synthesis, execution, and debug stages. SED first produces initial programs using the neural program synthesizer component, then utilizes a neural program debugger to iteratively repair the generated programs. On Karel, a challenging input-output program synthesis benchmark, SED reduces the error rate of the neural program synthesizer itself by a considerable margin, and outperforms the standard beam search for decoding.
  arXiv  Detail & Related papers  (2020-07-16T04:15:47Z)
• Towards Neural-Guided Program Synthesis for Linear Temporal Logic Specifications [26.547133495699093]
  We use a neural network to learn a Q-function that is then used to guide search, and to construct programs that are subsequently verified for correctness. Our method is unique in combining search with deep learning to realize synthesis.
  arXiv  Detail & Related papers  (2019-12-31T17:09:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.