Prime Agent: A Self-Improving RLM Harness

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 testtime 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, GPUkernel generation, emulator construction, and autonomous nanoGPT speedruns.

Introduction. A strong language model on its own does not have the full capabilities of a computer. An LLM is a bounded sequential processor whose next decision can use only state information exposed in its weights and active context. A harness supplies the missing computational substrate that allows for external actions via tool-calls. These external actions also include information management beyond prior knowledge stored in the model weights. Context management was first enabled by agentic compaction, the process by which a model selectively analyzes its own context to reduce tokens while keeping essential information. However, the full information state has grown beyond the weights and token context. Imagined as a state information cache (Figure 2), model weights are L0, active context is L1, a persistent REPL and recursive subagents form L2, and disk-backed history, memories, and skills form L3. This makes the system more von Neumann-like: the model can read, transform, and write addressable state outside the instruction currently being generated [34, 35].

Discussion / Conclusion. Prime Agent introduces a new paradigm for agent harness design in which persistent execution, recursive sessions, autonomous controls, recorded history, and Continual Harness form one substrate for long-horizon work. Results across interactive reasoning, long-context tasks, autonomous research, systems construction, and persistent environments show that this substrate supports different forms of test-time computation under standardized execution and accounting. Despite its results relative to alternative harnesses, models still experience friction when deciding how to allocate subagents, manage retained information, and refine reusable state. Many harness capabilities remain underused because current models were not trained to operate them. We expect model-harness co-learning to become the dominant route to new long-horizon capabilities.