Claude (design)vsKimi Agent
1 prompt where both agents ran the exact same instructions. Each row is one prompt; the figures are whatever was measured or reported for that run.
This page holds the harness constant, not the model: the program driving a model changes the result as much as the model does, so the two are compared separately.
Claude (design) and Kimi Agent ran the same prompt on 1 task, side by side. Cost, duration and outcome for each — 0 measured locally, no aggregate score.
- Shared prompts
- 1
- Compared
- coding agents
- Measured runs
- 0
No winner is declared. A measured run and a figure someone posted are not the same evidence, so they are never averaged into a ranking.
What this page shows, and how to read it
Claude (design) is an agentic coding tool from Anthropic, and Kimi Agent is an agentic coding tool from Moonshot AI. A harness is the program that receives the prompt and works the task — editing files, running commands — while a model decides each step. On the prompts below, Claude (design) drove Claude Opus 4.7, and Kimi Agent drove Kimi K2.6.
The two sides share 1 prompt on this site: "Sterrelicht — a stargazing lodge landing page, in one prompt". It is a websites prompt, rated hard, published by Fakhr on May 4, 2026, tagged landing-page, design, one-prompt and open-weights. Every run has a video capture on the prompt page. Across it, Claude (design) completed its run, and Kimi Agent completed its run. Claude (design) has no cost or duration on record for these runs. Kimi Agent has no cost or duration on record for these runs.
How to read the outcomes. Completed means the prompt alone produced a working artefact. Failed means this run did not — the build broke, the result would not run, or the agent stopped short of a working state. That is a fact about one run on one prompt, not a verdict on the tool: the same programs complete other prompts elsewhere on this site. Timed out means the run was cut at its time limit with the artefact unfinished, so whatever it cost bought a partial result. Stopped and Interrupted mark runs ended from the outside before they concluded. A low cost attached to a run that did not finish is not a saving — it is the price of an attempt, which is why every figure on this page travels with its status.
Every figure above carries a trust label. Measured means SamePrompt ran it locally in Bench Arena, reconciled the tokens on the harness log and recomputed the cost from them. Reported means the author of the source announced the figure; it is shown as stated and cannot be verified here. The two are never summed, averaged or ranked, and this page declares no winner: a cheap run that failed and an expensive run that completed are two facts, not a score.