Open-Source AI Instruction Library

One Prompt. Top Models. Scored.

The only library that measures the same instruction file across major models with deterministic, reproducible scoring.

Explore Library GitHub Repo
What makes it different
Multi-Format Exports Direct transpilation for Cursor Rules, Windsurf, Claude Projects, and raw system prompts.
Compatibility Verified Benchmark any file across 8 models from Anthropic, OpenAI, Google, Qwen, Zhipu, and MiniMax with transparent scoring.
Structured UIR Consistent metadata schema powering validation, transpilation, and compatibility scoring.
Latest Models · OpenRouter Our Roster →
~unbiased/pareto-latest New
Pareto Latest
2026-10-13 $2.50 in $7.50 out
unbiased/pareto-26.9 New
Pareto 26.9
2026-10-13 $2.50 in $7.50 out
liquid/lfm-2.5-24b-a2b New
LiquidAI: LFM2.5-24B-A2B
2026-10-13 $0.05 in $0.70 out
google/gemini-4-argon New
Google: Gemini 4 Argon
2026-10-13 $2.00 in $10.00 out
google/gemini-4-argon:batch New
Google: Gemini 4 Argon (batch)
2026-10-13 $1.00 in $5.00 out
scaledown/summarize New
ScaleDown: Summarize
2026-10-13 $0.04 in — out
scaledown/extract New
ScaleDown: Extract
2026-10-13 $0.04 in — out
meta/muse-wonder New
Meta: Muse Wonder
2026-10-12 $4.00 in $20.00 out
stepfun/step-5-preview New
StepFun: Step 5 Preview
2026-10-08 $1.00 in $2.70 out