curl https://selfbench.dev/api/public/results
{
"cards": [{
"repository": { "id": 1298096233, "fullName": "mupt-ai/self-bench" },
"publisher": { "login": "mupt-ai", "kind": "org" },
"releaseId": "c3779cc9-92c9-4111-83a9-8e6334138cc1",
"releasedAt": "2026-10-04T18:41:01.523Z",
"tasks": 38,
"settings": 6,
"picks": [{
"setting": { "id": "gpt-6-luna|codex|high", "model": { "label": "GPT-6 Luna" }, "accuracy": 47.37, "costPerTaskUsd": 0.02717 },
"roles": ["cheapest"]
}],
"frontier": [{ "id": "gpt-6-luna|codex|high", "model": { "label": "GPT-6 Luna" }, "accuracy": 47.37, "costPerTaskUsd": 0.02717 }],
"endorsed": false,
"defaultLine": true
}]
}
Public Results
List Results
List released repositories and leaderboard summaries.
GET
/
api
/
public
/
results
curl https://selfbench.dev/api/public/results
{
"cards": [{
"repository": { "id": 1298096233, "fullName": "mupt-ai/self-bench" },
"publisher": { "login": "mupt-ai", "kind": "org" },
"releaseId": "c3779cc9-92c9-4111-83a9-8e6334138cc1",
"releasedAt": "2026-10-04T18:41:01.523Z",
"tasks": 38,
"settings": 6,
"picks": [{
"setting": { "id": "gpt-6-luna|codex|high", "model": { "label": "GPT-6 Luna" }, "accuracy": 47.37, "costPerTaskUsd": 0.02717 },
"roles": ["cheapest"]
}],
"frontier": [{ "id": "gpt-6-luna|codex|high", "model": { "label": "GPT-6 Luna" }, "accuracy": 47.37, "costPerTaskUsd": 0.02717 }],
"endorsed": false,
"defaultLine": true
}]
}
Returns
Example abbreviated:
{ "cards": [...] }: one card per repository and publisher, newest first. defaultLine marks the publisher shown by default.
curl https://selfbench.dev/api/public/results
{
"cards": [{
"repository": { "id": 1298096233, "fullName": "mupt-ai/self-bench" },
"publisher": { "login": "mupt-ai", "kind": "org" },
"releaseId": "c3779cc9-92c9-4111-83a9-8e6334138cc1",
"releasedAt": "2026-10-04T18:41:01.523Z",
"tasks": 38,
"settings": 6,
"picks": [{
"setting": { "id": "gpt-6-luna|codex|high", "model": { "label": "GPT-6 Luna" }, "accuracy": 47.37, "costPerTaskUsd": 0.02717 },
"roles": ["cheapest"]
}],
"frontier": [{ "id": "gpt-6-luna|codex|high", "model": { "label": "GPT-6 Luna" }, "accuracy": 47.37, "costPerTaskUsd": 0.02717 }],
"endorsed": false,
"defaultLine": true
}]
}
picks highlights the cheapest and most accurate frontier settings; frontier lists all of them. accuracy is a percentage (0–100), and costPerTaskUsd is USD per task.