Vendor

All Moonshot AI AI models

Beijing-based vendor behind Kimi K2.7 — a frontier coder available on a flat plan.

Models on bench4
Total task attempts32
Scored cells14
Gold medals🥇 3
Vendor avg score7.25/10
Top modelKimi K2.7

My take on Moonshot AI

Moonshot AI's Kimi K2.7 is the sleeper hit of the bench — frontier-class coder on a flat-rate plan, which changes the economics of agent loops entirely. The three modes (Fast, No-Think, Quality) are useful in different ways: Fast for snappy iteration, Quality for one-shot polish, No-Think when the plan is already in the prompt. The flat plan means I can iterate without watching a meter.

Where I use Moonshot AI inside the Agent OS. Each model below has a "How I use it" line in its detail page — that's the daily-usage view, not the marketing pitch.

How I tested Moonshot AI's models

Every model on this page received the exact same fixed prompt as every other model on the bench. One shot, single HTML file out, scored 0–10 by me on three axes (runs, hits the brief, looks good). The scoring is published in my source comparison guides on agentos.guide — see the methodology page for full data provenance.

Vendor: moonshot.ai ↗

The same stack Julian uses

Run this stack yourself.

Every demo on this bench was built inside the Agent Operating System — one prompt, one shot, single HTML file out. The Agent OS, the prompts, the templates, the weekly walkthroughs and 3,600+ founders shipping with it every day all live inside the AI Profit Boardroom.

3,600+founders
258documented wins
38countries
$100k+/mocommunity MRR