analiza aiqbench
Prime Agent introduces a novel open-source framework that decouples long-horizon agent execution from model weights, enabling persistent state management and subagent coordination for complex coding workflows. By standardizing resource accounting and recovery mechanisms, this harness significantly improves evaluation metrics on benchmarks like ARC-AGI-3 while allowing researchers to isolate true model capabilities from infrastructure failures. The system demonstrates practical utility in parallelized tasks such as Factorio automation and GPU-kernel generation, marking a step forward in scalable agent architecture.
Prime Agent: A Self-Improving RLM Harness
arXiv:2608.23552v1 Announce Type: new Abstract: Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context. Prime Agent is an open-source harness for long-horizon evaluation and coding-agent workflows. A persistent IPython REPL follows the Recursive Language Model abstraction for programmatic context processing and test-time compute, while Continual Harness preserves histories, memories, skills, prompts, and subagent specifications across trajectories. Recursive subagents coordinate through direct agent-to-agent communication, and the Agents View lets humans inspect and manage daemon-backed sessions. Prime Agent standardizes execution, recovery, verification, and resource accounting while leaving strategy construction to the model. This low-friction, expressive membrane prevents harness failures from becoming model failures and pushes measurement toward the model's true maximal underlying capability. Prime Agent raises ARC-AGI-3 RHAE Best@1 from 30% to 95.5% and matches or exceeds native and popular harnesses across long-context coding, GPU-kernel generation, emulator construction, and autonomous nanoGPT speedruns. On Factorio, we find refinement allows for continuous technology progression and dedicated subagents enable parallelized work. Code is available at https://github.com/PrimeIntellect-ai/prime-agent.
Authors:Seth Karten, Alex L. Zhang, Kevin Thomas, Sebastian Müller, Elie Bakouch, Daniel Auras, Mika Senghaas, Fares Obeid, Konstantin Dunas, Johannes Hagemann, Sami Jaghouar
Abstract:Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context. Prime Agent is an open-source harness for long-horizon evaluation and coding-agent workflows. A persistent IPython REPL follows the Recursive Language Model abstraction for programmatic context processing and test-time compute, while Continual Harness preserves histories, memories, skills, prompts, and subagent specifications across trajectories. Recursive subagents coordinate through direct agent-to-agent communication, and the Agents View lets humans inspect and manage daemon-backed sessions. Prime Agent standardizes execution, recovery, verification, and resource accounting while leaving strategy construction to the model. This low-friction, expressive membrane prevents harness failures from becoming model failures and pushes measurement toward the model's true maximal underlying capability. Prime Agent raises ARC-AGI-3 RHAE Best@1 from 30% to 95.5% and matches or exceeds native and popular harnesses across long-context coding, GPU-kernel generation, emulator construction, and autonomous nanoGPT speedruns. On Factorio, we find refinement allows for continuous technology progression and dedicated subagents enable parallelized work. Code is available at this https URL.
16 pages, 10 figures. Technical report. Code: this https URL
Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Software Engineering (cs.SE)
arXiv:2608.23552 [cs.AI]
(or arXiv:2608.23552v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.23552 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 226 cuvinte
Subiecte conexe
[AINews] not much happened today
Latent Space · 18 sept. 2026
[AINews] not much happened today
Latent Space · 10 sept. 2026
Academia is for Ambition — Alex Zhang, MIT
Latent Space · 2 oct. 2026
Introducing North Mini Code: Cohere’s First Model For Developers
Hugging Face · 9 iun. 2026