catherd-cli 0.2.1 → 1.1.0

package/catalog/catalog.json

@@ -1,112 +0,0 @@
-{
-  "version": "2026-09-24",
-  "sources": {
-    "repo_code": "DeepSWE-style resolution of real repository issues, per model and effort, from the published results collected in spec §5 on 2026-09-24",
-    "terminal": "Terminal-Bench-style terminal, ops and environment work, per model and effort, from the same collection",
-    "honesty": "100 minus the rate of failing to disclose a broken tool, from the system cards: gpt-6-luna 71.3, gpt-6-sol 95.1, claude-opus-5-5 not published",
-    "secs_per_task": "median seconds per task in the same benchmark runs",
-    "costRank": "catherd's own order of what one run costs a user on a ChatGPT plan and a Claude plan, lowest first; subscription quota counts"
-  },
-  "notes": {
-    "gpt-6-sol": "context 400000 is unverified for the ChatGPT-login id",
-    "gpt-6-luna": "context 400000 is unverified for the ChatGPT-login id",
-    "claude-opus-5-5": "context 1000000 is unverified",
-    "unscored on purpose": "gpt-6-luna#xhigh was dominated; gpt-6-luna#max took 958 s a task, slower than gpt-6-sol#xhigh at an equal score"
-  },
-  "models": [
-    {
-      "id": "gpt-6-luna",
-      "backend": "codex",
-      "efforts": ["low", "medium", "high", "xhigh", "max"],
-      "capabilities": {
-        "toolCall": true,
-        "imageIn": true,
-        "imageOut": true,
-        "reasoning": true,
-        "context": 400000
-      }
-    },
-    {
-      "id": "gpt-6-sol",
-      "backend": "codex",
-      "efforts": ["low", "medium", "high", "xhigh"],
-      "capabilities": {
-        "toolCall": true,
-        "imageIn": true,
-        "imageOut": true,
-        "reasoning": true,
-        "context": 400000
-      }
-    },
-    {
-      "id": "claude-opus-5-5",
-      "backend": "claude",
-      "efforts": ["low", "medium", "high"],
-      "capabilities": {
-        "toolCall": true,
-        "imageIn": true,
-        "imageOut": false,
-        "reasoning": true,
-        "context": 1000000
-      }
-    }
-  ],
-  "entries": [
-    {
-      "rung": "gpt-6-luna#high",
-      "scores": { "repo_code": 59.3, "terminal": 4.5, "honesty": 71.3, "secs_per_task": 563 },
-      "costRank": 1
-    },
-    {
-      "rung": "gpt-6-sol#medium",
-      "scores": { "repo_code": 56.6, "terminal": 18.7, "honesty": 95.1, "secs_per_task": 264 },
-      "costRank": 2
-    },
-    {
-      "rung": "gpt-6-sol#high",
-      "scores": { "repo_code": 65.3, "terminal": 26.3, "honesty": 95.1, "secs_per_task": 391 },
-      "costRank": 3
-    },
-    {
-      "rung": "gpt-6-sol#xhigh",
-      "scores": { "repo_code": 66.6, "terminal": 30.3, "honesty": 95.1, "secs_per_task": 496 },
-      "costRank": 4
-    },
-    { "rung": "claude-opus-5-5#low", "scores": { "terminal": 31.3 }, "costRank": 5 },
-    { "rung": "claude-opus-5-5#medium", "scores": { "terminal": 52.5 }, "costRank": 6 },
-    { "rung": "claude-opus-5-5#high", "scores": { "terminal": 56.6 }, "costRank": 7 }
-  ],
-  "bars": {
-    "repo_code": {
-      "copy": { "repo_code": 55 },
-      "build": { "repo_code": 55 },
-      "logic": { "repo_code": 55, "terminal": 15 },
-      "hard": { "repo_code": 56, "terminal": 18 }
-    },
-    "terminal": {
-      "copy": { "terminal": 15 },
-      "build": { "terminal": 15 },
-      "logic": { "terminal": 15 },
-      "hard": { "terminal": 15 }
-    },
-    "ui": {
-      "copy": { "repo_code": 55 },
-      "build": { "repo_code": 55 },
-      "logic": { "repo_code": 55, "terminal": 15 },
-      "hard": { "repo_code": 56, "terminal": 18 }
-    },
-    "prose": {
-      "copy": { "repo_code": 55 },
-      "build": { "repo_code": 55 },
-      "logic": { "repo_code": 55, "terminal": 15 },
-      "hard": { "repo_code": 56, "terminal": 18 }
-    },
-    "research": {
-      "copy": { "repo_code": 55 },
-      "build": { "repo_code": 55 },
-      "logic": { "repo_code": 55, "terminal": 15 },
-      "hard": { "repo_code": 56, "terminal": 18 }
-    }
-  },
-  "treatLike": {}
-}