Skip to content

docs(benchmarking): Tighten overview analysis #421

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
dcramer merged 1 commit into from
Jul 2, 2026
Merged

docs(benchmarking): Tighten overview analysis #421

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
128 changes: 49 additions & 79 deletions packages/docs/src/components/BenchmarkRuns.astro
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,34 @@ type BenchmarkResult = {
};
};

const MODEL_LABELS: Record<string, string> = {
"anthropic/claude-opus-4-6": "Claude Opus 4.6",
"anthropic/claude-opus-4-7": "Claude Opus 4.7",
"anthropic/claude-opus-4-8": "Claude Opus 4.8",
"anthropic/claude-sonnet-4-6": "Claude Sonnet 4.6",
"anthropic/claude-sonnet-5": "Claude Sonnet 5",
"claude-opus-4-6": "Claude Opus 4.6",
"claude-opus-4-7": "Claude Opus 4.7",
"claude-opus-4-8": "Claude Opus 4.8",
"claude-sonnet-4-6": "Claude Sonnet 4.6",
"openai/gpt-5.5": "GPT 5.5",
"openrouter/deepseek/deepseek-v4-flash": "DeepSeek V4 Flash",
"openrouter/deepseek/deepseek-v4-pro": "DeepSeek V4 Pro",
};

const REASONING_LABELS: Record<string, string> = {
high: "high",
low: "low",
"pi-default-medium": "medium",
"provider-default": "default",
xhigh: "xhigh",
};

const RUNTIME_LABELS: Record<string, string> = {
claude: "Claude SDK",
pi: "Pi",
};

const knownFoundRate = (scoring: BenchmarkScoring | undefined) =>
scoring && scoring.knownFindingCount !== 0
? scoring.knownFound / scoring.knownFindingCount
Expand Down Expand Up @@ -79,74 +107,12 @@ const formatDuration = (value: number | undefined) => {

const formatPercent = (value: number) => `${(value * 100).toFixed(1)}%`;

const modelLabel = (model: string) => {
if (model === "openai/gpt-5.5") {
return "GPT 5.5";
}
if (model === "openrouter/deepseek/deepseek-v4-pro") {
return "DeepSeek V4 Pro";
}
if (model === "openrouter/deepseek/deepseek-v4-flash") {
return "DeepSeek V4 Flash";
}
if (model === "anthropic/claude-sonnet-4-6") {
return "Claude Sonnet 4.6";
}
if (model === "anthropic/claude-sonnet-5") {
return "Claude Sonnet 5";
}
if (model === "claude-sonnet-4-6") {
return "Claude Sonnet 4.6";
}
if (model === "anthropic/claude-opus-4-8") {
return "Claude Opus 4.8";
}
if (model === "claude-opus-4-8") {
return "Claude Opus 4.8";
}
if (model === "anthropic/claude-opus-4-6") {
return "Claude Opus 4.6";
}
if (model === "claude-opus-4-6") {
return "Claude Opus 4.6";
}
if (model === "anthropic/claude-opus-4-7") {
return "Claude Opus 4.7";
}
if (model === "claude-opus-4-7") {
return "Claude Opus 4.7";
}
return model;
};
const modelLabel = (model: string) => MODEL_LABELS[model] ?? model;

const runtimeLabel = (runtime: string) => {
if (runtime === "claude") {
return "Claude SDK";
}
if (runtime === "pi") {
return "Pi";
}
return runtime;
};
const runtimeLabel = (runtime: string) => RUNTIME_LABELS[runtime] ?? runtime;

const reasoningLabel = (reasoningLevel: string | undefined) => {
if (reasoningLevel === "pi-default-medium") {
return "medium";
}
if (reasoningLevel === "provider-default") {
return "default";
}
if (reasoningLevel === "low") {
return "low";
}
if (reasoningLevel === "high") {
return "high";
}
if (reasoningLevel === "xhigh") {
return "xhigh";
}
return undefined;
};
const reasoningLabel = (reasoningLevel: string | undefined) =>
reasoningLevel ? REASONING_LABELS[reasoningLevel] : undefined;

const hidesReasoningLabel = (result: BenchmarkResult) =>
result.model.includes("claude-sonnet");
Expand Down Expand Up @@ -213,12 +179,16 @@ const compareTimingResults = (a: BenchmarkResult, b: BenchmarkResult) => {
const aTiming = a.timing?.analysisChunkMs;
const bTiming = b.timing?.analysisChunkMs;

const p50Diff = (aTiming?.p50 ?? Number.POSITIVE_INFINITY) - (bTiming?.p50 ?? Number.POSITIVE_INFINITY);
const p50Diff =
(aTiming?.p50 ?? Number.POSITIVE_INFINITY) -
(bTiming?.p50 ?? Number.POSITIVE_INFINITY);
if (p50Diff !== 0) {
return p50Diff;
}

const p90Diff = (aTiming?.p90 ?? Number.POSITIVE_INFINITY) - (bTiming?.p90 ?? Number.POSITIVE_INFINITY);
const p90Diff =
(aTiming?.p90 ?? Number.POSITIVE_INFINITY) -
(bTiming?.p90 ?? Number.POSITIVE_INFINITY);
if (p90Diff !== 0) {
return p90Diff;
}
Expand Down Expand Up @@ -247,7 +217,7 @@ const timingResults = [...stableResults].sort(compareTimingResults);
<span role="columnheader">Run</span>
<span role="columnheader">Known</span>
<span role="columnheader">Findings</span>
<span role="columnheader">Recorded Cost</span>
<span role="columnheader">Cost</span>
</div>
{scoreResults.map((result) => {
const scoring = result.scoring;
Expand Down Expand Up @@ -275,7 +245,7 @@ const timingResults = [...stableResults].sort(compareTimingResults);
<strong class="metric-value">{result.summary.findingsTotal}</strong>
</div>
<div class="metric-cell" role="cell">
<span class="cell-label">Recorded cost</span>
<span class="cell-label">Cost</span>
<strong class="metric-value">{formatCost(result.summary.costUSD)}</strong>
</div>
</article>
Expand All @@ -287,15 +257,15 @@ const timingResults = [...stableResults].sort(compareTimingResults);
{stableResults.length > 0 && (
<div class="detail-tables">
<section class="detail-section" aria-labelledby="benchmark-cost-breakdown">
<h3 id="benchmark-cost-breakdown">Cost and Tokens</h3>
<p class="sort-note">Sorted by recorded cost, lowest first.</p>
<h3 id="benchmark-cost-breakdown">Cost</h3>
<p class="sort-note">Lowest cost first.</p>
<div class="detail-table-wrap">
<div class="run-table cost-table" role="table" aria-label="Cost breakdown">
<div class="run-header" role="row">
<span role="columnheader">Run</span>
<span role="columnheader">Recorded Cost</span>
<span role="columnheader">Input Tokens</span>
<span role="columnheader">Output Tokens</span>
<span role="columnheader">Cost</span>
<span role="columnheader">Input</span>
<span role="columnheader">Output</span>
</div>
{costResults.map((result) => (
<article class="run-row" role="row">
Expand All @@ -307,15 +277,15 @@ const timingResults = [...stableResults].sort(compareTimingResults);
{runMeta(result) && <p>{runMeta(result)}</p>}
</div>
<div class="metric-cell" role="cell">
<span class="cell-label">Recorded cost</span>
<span class="cell-label">Cost</span>
<strong class="metric-value">{formatCost(result.summary.costUSD)}</strong>
</div>
<div class="metric-cell" role="cell">
<span class="cell-label">Input tokens</span>
<span class="cell-label">Input</span>
<strong class="metric-value">{formatTokenCount(result.summary.inputTokens)}</strong>
</div>
<div class="metric-cell" role="cell">
<span class="cell-label">Output tokens</span>
<span class="cell-label">Output</span>
<strong class="metric-value">{formatTokenCount(result.summary.outputTokens)}</strong>
</div>
</article>
Expand All @@ -326,7 +296,7 @@ const timingResults = [...stableResults].sort(compareTimingResults);

<section class="detail-section" aria-labelledby="benchmark-timing-breakdown">
<h3 id="benchmark-timing-breakdown">Timing</h3>
<p class="sort-note">Sorted by P50 analysis chunk duration, lowest first.</p>
<p class="sort-note">Lowest P50 chunk duration first.</p>
<div class="detail-table-wrap">
<div class="run-table timing-table" role="table" aria-label="Timing breakdown">
<div class="run-header" role="row">
Expand Down
Loading
Loading