Noesion Radarnoesion.ai
Research2026-10-09

Vals ran GPT 6 Sol and Claude Opus 5.5 on Vibe Code Bench as single agents and as teams. Teams cost 1.8 to 5.1 times as much. Only Sol at medium effort showed a statistically significant score gain. Sol’s lead split work early and tested itself. Opus’s lead wrote a contract first and delegated in waves. Sol teams finished faster at both efforts. Opus teams took longer at max effort. Higher effort raised Sol’s score more than a team did. Opus’s cheapest setup scored about as well as its most expensive one. Findings may not apply to other kinds of work.

What changed

Single agents handled entire tasks without delegation.

What it unlocks

Splitting work across parallel subagents to build and test parts of an application.

  • Teams cost 1.8 to 5.1 times as much as single agents
  • Sol medium-effort team +7.3 points (p = 0.005)
  • Opus max-effort team cost $122 per app
  • Opus subagents read 224M cached tokens per app at max effort

Sources

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.