Set the percentage of tests to run during the benchmark. 100% runs all tests.

<!-- CONFIG_START -->
RUN_PERCENTAGE: 100
SHARED_PROMPT: "Provide production-ready and maintainable JavaScript code. Apply code golfing practices but don't put everything in a single line. No comments. Your code will execute in the browser."
<!-- CONFIG_END -->


The following models are included in the benchmark run.

<!-- MODELS_START -->
google/gemini-3-pro-preview
anthropic/claude-sonnet-4.5 TEMP:0.7
openai/gpt-5.1-codex
moonshotai/kimi-k2-thinking
google/gemini-2.5-pro
openrouter/sherlock-think-alpha
<!-- MODELS_END -->

The following will be available on lynchmark.com/gemini
We will use google api instead of openrouter using secret GOOGLE_KEY

<!-- GEMINI_START -->
gemini-3-pro-preview TEMP:2
gemini-3-pro-preview TEMP:1.9
gemini-3-pro-preview TEMP:1.8
gemini-3-pro-preview TEMP:1.7
gemini-3-pro-preview TEMP:1.6
gemini-3-pro-preview TEMP:1.5
gemini-3-pro-preview TEMP:1.4
gemini-3-pro-preview TEMP:1.3
gemini-3-pro-preview TEMP:1.2
gemini-3-pro-preview TEMP:1.1
gemini-3-pro-preview TEMP:1
gemini-3-pro-preview TEMP:0.9
gemini-3-pro-preview TEMP:0.8
gemini-3-pro-preview TEMP:0.7
gemini-3-pro-preview TEMP:0.6
gemini-3-pro-preview TEMP:0.5
gemini-3-pro-preview TEMP:0.4
gemini-3-pro-preview TEMP:0.3
gemini-3-pro-preview TEMP:0.2
gemini-3-pro-preview TEMP:0.1
gemini-3-pro-preview TEMP:0
<!-- MODELS_END -->