Terminal-Bench: where does Little Coder actually land?

“Beats Gemini 2.5 + Gemini CLI” is true — but Gemini 2.5 is last generation. Here's the same number with the rest of the field on the chart. Share of tasks solved, higher is better.

The blue bar runs on a laptop — free, offline, no rate limits. Every gray bar is a metered cloud API. Source: Terminal-Bench 2.0 leaderboard (tbench.ai), July 2026. Each score ±2–4pp.