GoKawiilUnreal Labs launches Unreal Agent, an agent harness cutting costs vs Codex
Unreal Labs introduced Unreal Agent, a harness for running AI agents that it says delivers up to 40% cost savings compared to OpenAI's Codex and up to 20% savings compared to Pi, without hurting performance on production workloads and coding/science benchmarks. The company says the harness manages tool calls asynchronously, removing the need for the underlying model to handle waits, polls, and heartbeats itself.
GoKawiil's interpretation of the reporting above, not reported fact.
By offloading tool-call management from the model, Unreal Labs suggests agents could respond faster to users while packing more useful work between model calls, which the company frames as a route to lower operating costs at scale. The claims about savings and benchmark performance come solely from Unreal Labs itself, so independent verification of the figures is not yet available. The move also signals growing competition around agent harness design as a distinct layer separate from the underlying language models.
- Unreal Agent claims up to 40% cost savings versus Codex and 20% versus Pi on benchmarks
- The harness handles tool calls asynchronously, reducing model overhead for waits and polling
- Unreal Labs is soliciting partnerships for cost-efficient agent deployment via [email protected]
Source: unreallabs.ai, 2026-09-22
Published there as: “Unreal Agent”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.