Skip to content

Make Codex and Claude Code subscriptions go further: a local model router that gives routine agent work to economical models and keeps your strongest model for the hard parts.

725 B
Raw Blame
{
  "id": "format-development-completed",
  "dataset": "engineering",
  "kind": "format_development",
  "repeats": 1,
  "planned_calls": 20,
  "result_calls": 20,
  "ledger_scope": "Original result-only developer ledger; no start events recorded. Unobserved plan entries are not imputed as observed starts. Not a headline benchmark.",
  "unobserved_plan_entries": 0,
  "expected_models": {
    "qwen38flash": "qwen3.8-flash",
    "jev113": "typesafe/jev-1.13-20260917",
    "jev114": "typesafe/jev-1.14-20261014"
  },
  "jobs": "execution/format-development-completed/jobs.jsonl.gz",
  "requests": "execution/format-development-completed/requests.jsonl.gz",
  "responses": "responses/format-development-completed.jsonl.gz"
}