{"slug": "llm-as-a-verifier-a-general-purpose-verification-framework", "title": "LLM-as-a-Verifier: A General-Purpose Verification Framework", "summary": "Researchers introduced LLM-as-a-Verifier, a general-purpose verification framework that generates continuous scores from LLM logits to assess solution correctness without additional training. The framework achieved state-of-the-art results on multiple benchmarks, including Terminal-Bench V2 (86.5%), SWE-Bench Verified (78.2%), RoboRewardBench (87.4%), and MedAgentBench (73.3%), and can provide dense feedback for reinforcement learning.", "body_md": "# Computer Science > Artificial Intelligence\n\n[Submitted on 6 Jul 2026 (\n\n[v1](https://arxiv.org/abs/2607.05391v1)), last revised 7 Jul 2026 (this version, v2)]# Title:LLM-as-a-Verifier: A General-Purpose Verification Framework\n\n[View PDF](/pdf/2607.05391)\n\n[HTML (experimental)](https://arxiv.org/html/2607.05391v2)\n\nAbstract:Scaling pre-training, post-training, and test-time compute have become the central paradigms for improving the capabilities of LLMs. In this work, we identify verification, the ability to determine the correctness of a solution, as a new scaling axis. To unlock this and demonstrate its effectiveness, we introduce LLM-as-a-Verifier, a general-purpose verification framework that provides fine-grained feedback for agentic tasks without requiring additional training. Unlike standard LM judges that prompt LLMs to produce discrete scores for candidate solutions, LLM-as-a-Verifier computes the expectation over the distribution of scoring token logits to generate continuous scores. This probabilistic formulation enables verification to scale along multiple dimensions: (1) score granularity, (2) repeated evaluation, and (3) criteria decomposition. In particular, we show that scaling the scoring granularity leads to better separation between positive and negative solutions, resulting in more calibrated comparisons. Moreover, scaling repeated evaluation and criteria decomposition consistently lead to additional gains in verification accuracy through variance and complexity reduction. We further introduce a cost-efficient ranking algorithm for selecting the best solution among candidates using the verifier's continuous scores. LLM-as-a-Verifier achieves state-of-the-art performance on Terminal-Bench V2 (86.5%), SWE-Bench Verified (78.2%), RoboRewardBench (87.4%), and MedAgentBench (73.3%). Beyond verification, the fine-grained signals from LLM-as-a-Verifier can also serve as a proxy for estimating task progress. We build an extension for Claude Code, enabling developers to monitor and improve their own agentic systems. Finally, we show that LLM-as-a-Verifier can provide dense feedback for RL, improving the sample efficiency of SAC and GRPO on robotics and mathematical reasoning benchmarks.\n\n## Submission history\n\nFrom: Jacky Kwok [[view email](/show-email/99fef068/2607.05391)]\n\n**Mon, 6 Jul 2026 17:59:35 UTC (4,216 KB)**\n\n[[v1]](/abs/2607.05391v1)**[v2]** Tue, 7 Jul 2026 17:26:37 UTC (4,211 KB)\n\n### Current browse context:\n\ncs.AI\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer\n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers\n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps\n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations\n\n*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article\n\nalphaXiv\n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers\n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub\n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub\n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face\n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast\n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower\n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender\n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/llm-as-a-verifier-a-general-purpose-verification-framework", "canonical_source": "https://arxiv.org/abs/2607.05391", "published_at": "2026-07-14 23:09:17+00:00", "updated_at": "2026-07-14 23:17:54.648264+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "ai-agents"], "entities": ["LLM-as-a-Verifier", "Terminal-Bench V2", "SWE-Bench Verified", "RoboRewardBench", "MedAgentBench", "Claude Code", "SAC", "GRPO"], "alternates": {"html": "https://wpnews.pro/news/llm-as-a-verifier-a-general-purpose-verification-framework", "markdown": "https://wpnews.pro/news/llm-as-a-verifier-a-general-purpose-verification-framework.md", "text": "https://wpnews.pro/news/llm-as-a-verifier-a-general-purpose-verification-framework.txt", "jsonld": "https://wpnews.pro/news/llm-as-a-verifier-a-general-purpose-verification-framework.jsonld"}}