{"slug": "hybrid-reasoning-agents-35-55-fewer-tokens-and-up-to-2-6x-faster-inference", "title": "Hybrid Reasoning Agents: 35–55% fewer tokens and up to 2.6× faster inference", "summary": "A paper submitted to arXiv on 13 October 2026 introduces Hybrid Reasoning Agents (HybridAgents), which switch between latent reasoning and text-based interaction during a single generation process, improving performance by 10.4% over frontier LLM agents and agentic frameworks across 11 benchmarks while cutting tokens by 35.5%–54.7%, speeding inference 1.8x–2.6x and reducing cost 2.7x–3.7x. The architecture uses a lightweight HybridLink to move between text and latent modes and a continuous working memory that preserves lossless hybrid context across turns, with HybridSFT and HybridRL post-training teaching the backbone when to think in latent space and when to act through text. The authors evaluated HybridAgent at 3 model scales on benchmarks covering mathematics, science, medicine, search and agentic reasoning, and released code and data at github.com/togethercomputer/HybridAgent.", "body_md": "# Computer Science > Computation and Language\n\n  [Submitted on 13 Oct 2026]\n\n# Title:Hybrid Reasoning Agents\n\n[View PDF](https://arxiv.org/pdf/2610.18861)\n\n[HTML (experimental)](https://arxiv.org/html/2610.18861v1)\n\nAbstract:How can we build an agent that thinks beyond human language, yet still acts through it in the external environment? Modern LLM agents still reason and interact through text, while latent reasoning offers an efficient internal medium but lacks native tool access and reasoning transparency. We bridge these two directions with Hybrid Reasoning Agents (HybridAgents), which dynamically navigate between latent reasoning and text-based interaction throughout a unified generation process. Architecturally, a lightweight HybridLink enables seamless transitions between text and latent modes, while the agent's continuous working memory preserves lossless hybrid context across multiple turns. These designs preserve tool-integrated agentic capabilities while making latent thoughts directly inspectable. Building on this architecture, we further develop dedicated post-training algorithms for hybrid reasoning: HybridSFT stabilizes tool-integrated hybrid generation by teaching the backbone to coordinate with HybridLink, then HybridRL scales reasoning by learning when to think in latent space and when to act through text. We evaluate HybridAgent at 3 model scales across 11 benchmarks covering mathematics, science, medicine, search, and agentic reasoning. Compared with frontier LLM Agents and agentic frameworks spanning both text-based and latent methods, HybridAgent consistently improves performance by 10.4%, while delivering 35.5%-54.7% token reduction, 1.8x-2.6x faster inference, and 2.7x-3.7x cost savings. Code and data are available at [this https URL](https://github.com/togethercomputer/HybridAgent).\n    \n\n### Additional Features\n\n### Current browse context:\n\ncs.CL\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/))\n# 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))\n# 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))\n# 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/hybrid-reasoning-agents-35-55-fewer-tokens-and-up-to-2-6x-faster-inference", "canonical_source": "https://arxiv.org/abs/2610.18861", "published_at": "2026-10-18 20:48:14+00:00", "updated_at": "2026-10-18 21:18:43.687074+00:00", "lang": "en", "topics": ["ai-agents", "ai-research", "large-language-models", "machine-learning", "artificial-intelligence"], "entities": ["HybridAgents", "HybridLink", "HybridSFT", "HybridRL", "arXiv", "Together Computer"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/hybrid-reasoning-agents-35-55-fewer-tokens-and-up-to-2-6x-faster-inference", "markdown": "https://wpnews.pro/news/hybrid-reasoning-agents-35-55-fewer-tokens-and-up-to-2-6x-faster-inference.md", "text": "https://wpnews.pro/news/hybrid-reasoning-agents-35-55-fewer-tokens-and-up-to-2-6x-faster-inference.txt", "jsonld": "https://wpnews.pro/news/hybrid-reasoning-agents-35-55-fewer-tokens-and-up-to-2-6x-faster-inference.jsonld"}}