Claude Haiku 5.5 (effort default (no --effort flag) · thinking on) on Model ladder: fix 10 real merged bugs with tests hidden

[
 {
  "duration_api_ms": 30882,
  "stop_reason": "end_turn",
  "session_id": "[REDACTED]",
  "total_cost_usd": 0.00904065,
  "usage": {
   "input_tokens": 18,
   "cache_creation_input_tokens": 18591,
   "cache_read_input_tokens": 218265,
   "output_tokens": 6276,
   "output_tokens_details": {
    "thinking_tokens": 3564
   },
   "server_tool_use": {
    "web_search_requests": 0,
    "web_fetch_requests": 0
   },
   "service_tier": "standard",
   "cache_creation": {
    "ephemeral_1h_input_tokens": 18591,
    "ephemeral_5m_input_tokens": 0
   },
   "inference_geo": "global",
   "iterations": [
    {
     "input_tokens": 2,
     "output_tokens": 532,
     "cache_read_input_tokens": 30553,
     "cache_creation_input_tokens": 577,
     "cache_creation": {
      "ephemeral_5m_input_tokens": 0,
      "ephemeral_1h_input_tokens": 577
     },
     "type": "message"
    }
   ],
   "speed": "standard",
   "fallback_credit": null
  },
  "modelUsage": {
   "claude-haiku-5-5": {
    "inputTokens": 18,
    "outputTokens": 6276,
    "cacheReadInputTokens": 218265,
    "cacheCreationInputTokens": 18591,
    "webSearchRequests": 0,
    "costUSD": 0.00904065,
    "contextWindow": 1000000,
    "maxOutputTokens": 128000,
    "thinkingTokens": 3564,
    "canonicalModel": "claude-haiku-5-5",
    "provider": "firstParty",
    "costBasis": "list"
   }
  },
  "permission_denials": [],
  "terminal_reason": "completed",
  "fast_mode_state": "off",
  "fast_mode_disabled_reason": "sdk_opt_in_required",
  "subagent_stats": {
   "spawned": 0,
   "requested": {
    "background": 0,
    "foreground": 0,
    "unset": 0
   },
   "started_in_background": 0,
   "max_depth": 0,
   "spawned_by_subagents": 0,
   "completed": 0,
   "failed": 0,
   "killed": {
    "parent": 0,
    "user": 0,
    "system": 0
   },
   "refused": {
    "depth_limit": 0,
    "concurrency_limit": 0,
    "budget": 0
   },
   "by_type": {}
  },
  "safety_stops": 0,
  "is_error": false,
  "num_turns": 9,
  "subtype": "success",
  "api_error_status": null,
  "result": "`intcomma()` now returns full, correctly grouped output for huge integers, both positive and negative, instead of raising `OverflowError`.\n\n**Cause:** `intcomma` called `float(value)` on every non-string input just to check for NaN/inf. For ints beyond float range, that raises `OverflowError`. Strings had a related bug: `float(\"1\"*400)` quietly returns `inf`, so `intcomma(\"1\"*400)` produced `+Inf`.\n\n**Fix** (`src/humanize/number.py`, `intcomma`):\n- Integer strings (optional sign, digits only) are parsed with `int()` directly.\n- Non-string inputs skip the float check when they're already `int`. Floats and other types keep the finiteness check, so NaN/inf and `None` behave as before.\n- Non-integer strings keep the existing path, so `\"nan\"` \u2192 `NaN`, `\"1e5\"` \u2192 `1e5`, and `\"abc\"` \u2192 `abc` are unchanged.\n\n**Tests** (`tests/test_number.py`): added cases for `10**400 + 123` and its negative, both as int and as string.\n\n**Results:**\n- `intcomma(10**400+123)` and `intcomma(-(10**400+123))` both end in `00,000,123`.\n- `.venv/bin/pytest --ignore=tests/test_benchmarks.py`: 713 passed, 110 skipped.\n- The full run also reports 15 errors in `tests/test_benchmarks.py`. These happen without my change too (checked with `git stash`); the `benchmark` fixture isn't available in this environment, so I left them alone.",
  "ttft_ms": 1626,
  "type": "result",
  "duration_ms": 34998,
  "uuid": "[REDACTED]",
  "ttft_stream_ms": 1193,
  "time_to_request_ms": 215,
  "first_content_frame_ms": 1193,
  "queued_turn_count": 0,
  "result_index": 0
 }
]