LoreonLabsPlatform

Overview

Intelligence

Markets
Builders
Research
Narratives
Ecosystems
Launchpads

Discover

Search
Sources

Python

kimi-tools

Kimi K2 (Moonshot) system prompts for coding agents + an A/B benchmark harness (cline/opencode/pi) and SWE-bench Verified eval

PythonEmerging

Stars

—

Forks

—

Contributors

1

Last push

4h ago

Recent commits

Latest commits.

ab: add DeepSeek-V4-Pro arm (deepseek4) to MODELS
7869364Fabian4h ago
docs: reframe as compliance-not-capability; add attempt-rate decomposition
c9adaecFabian4h ago
README: add Opus-4.8-xhigh to the headline table + charts
a154b9fFabian5h ago
README: rebase on the 48-instance harder band; retire the noisy /8 headline
cb020a2Fabian6h ago
ab: add Opus 4.8 xhigh×cursor probe (35/43); prompt sensitivity is flat
fc3feb4Fabian6h ago

ab: add Opus 4.8 arms (high/xhigh via --variant); xhigh×claude probe = 36/43

610e0d9Fabian7h ago

ab: full GLM-5.2 6-prompt sweep on harder band; prompt effect is model-specific

8f56e96Fabian19h ago

ab: GLM-5.2 third model on harder band + lead with 48-instance results

b267716Fabian1d ago

Top contributors

Builders behind this project.