Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions benchmarks/generate-supergpqa-chart.mjs
Original file line number Diff line number Diff line change
Expand Up @@ -236,9 +236,9 @@ function renderSvg(aggregate) {
lines.push(" <g font-family=\"Inter, ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, Segoe UI, sans-serif\">");
lines.push(` <text x="${plot.left}" y="31" fill="#111827" font-size="20" font-weight="650">SuperGPQA decision benchmark</text>`);
lines.push(` <text x="${plot.left}" y="52" fill="#6b7280" font-size="12">proportional sample · ${aggregate.dataset.rows} questions · ChooseKit OpenRouter providers pinned · 2026-09-21</text>`);
lines.push(` <line x1="548" y1="29" x2="558" y2="29" stroke="#15966b" stroke-width="1.5"/><text x="566" y="33" fill="#4b5563" font-size="11">ChooseKit + llama.cpp</text>`);
lines.push(` <circle cx="708" cy="29" r="5" fill="#356ae6"/><text x="720" y="33" fill="#4b5563" font-size="11">ChooseKit + OpenRouter</text>`);
lines.push(` <path d="M 874 23 L 880 29 L 874 35 L 868 29 Z" fill="#e85d3f"/><text x="888" y="33" fill="#4b5563" font-size="11">Jev</text>`);
lines.push(` <line x1="490" y1="29" x2="500" y2="29" stroke="#15966b" stroke-width="1.5"/><text x="508" y="33" fill="#4b5563" font-size="10">ChooseKit + llama.cpp</text>`);
lines.push(` <circle cx="650" cy="29" r="5" fill="#356ae6"/><text x="662" y="33" fill="#4b5563" font-size="10">ChooseKit + OpenRouter</text>`);
lines.push(` <path d="M 816 23 L 822 29 L 816 35 L 810 29 Z" fill="#e85d3f"/><text x="830" y="33" fill="#4b5563" font-size="10">Jev</text>`);

for (const tick of yTicks) {
const y = scaleY(tick);
Expand Down
6 changes: 3 additions & 3 deletions benchmarks/supergpqa-benchmark.svg
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading