gpt-5.1 @ low · cursor-xl

cursor-xllow
config
gpt-5.1-eff-low--cli--cursor-xl--dev
batch
2026-07-19--unified
transport
cursor-agent-cli-isolated-home
protocol
cli · plan
采集于
2026年7月18日 11:54 · N=2
该 model 的全部 effort 与 arm →
variant
live
slot 0 · P-q · sandboxed iframe
该 slot 没有截图

截图是 similarity 通道测量时的原始像素。运行 live 会在 sandboxed iframe 中重新执行原始输出;渲染结果取决于你的浏览器和字体,不保证与测量像素一致。

similarity 邻居本 session 没有 similarity 数据 —— 这个 batch 没有计算 pairs。

没有 valid 输出

这个 config 的每个 slot 都渲染失败了 —— 失败本身是数据,不是被藏起来的空缺。下方的 state 芯片与 send-log 会精确显示发生了什么。

Fidelity —— 卡片展示的数据对不对?

没有 fidelity 判定 —— 该 slot 为 infra-failed(fidelity 仅对 valid slot 计算)。

源码 —— 原始模型输出,一字节未改

该 slot 没有 card.html(infra-failed)。

L1 —— 廉价的确定性描述量

没有 L1 —— 该 slot 为 infra-failed。

该 slot 的 telemetry

prompt tokens0completion tokens0
total tokens0wall5.9 s
cost—request id—
config.json542 B
{
  "N": 2,
  "auth": {
    "credential_ref": "WCB_NONE",
    "method": "oauth-login"
  },
  "billing": "plan",
  "config_id": "gpt-5.1-eff-low--cli--cursor-xl--dev",
  "effort": "low",
  "family": "gpt",
  "m": {
    "min": 0,
    "q": 0
  },
  "model_id": "gpt-5.1",
  "protocol": "cli",
  "served_model": null,
  "telemetry": {
    "completion_tokens": null,
    "cost_usd": null,
    "prompt_tokens": null,
    "total_tokens": null,
    "wall_ms": null
  },
  "transport": "cursor-agent-cli-isolated-home",
  "vendor_sanction_ref": null
}
send-log2.6 KB
{
  "N": 2,
  "batch_id": "2026-08-13--cursor-xl-s1",
  "config_id": "gpt-5.1-eff-low--cli--cursor-xl--dev",
  "positions": [
    {
      "attempts": [
        {
          "attempt_index": 0,
          "backoff_ms": null,
          "charged": true,
          "ended_at": "2026-08-13T16:56:13.816233+00:00",
          "http_status": null,
          "outcome": "infra-failed",
          "reached_model": true,
          "reason": "mid-stream-death",
          "request_id": null,
          "started_at": "2026-08-13T16:56:07.911342+00:00"
        }
      ],
      "block_index": 0,
      "model_reaching_attempt_index": 0,
      "slot_index": 0,
      "terminal_state": "infra-failed",
      "variant": "P-q"
    },
    {
      "attempts": [
        {
          "attempt_index": 0,
          "backoff_ms": null,
          "charged": true,
          "ended_at": "2026-08-13T16:56:20.140081+00:00",
          "http_status": null,
          "outcome": "infra-failed",
          "reached_model": true,
          "reason": "mid-stream-death",
          "request_id": null,
          "started_at": "2026-08-13T16:56:13.818333+00:00"
        }
      ],
      "block_index": 0,
      "model_reaching_attempt_index": 0,
      "slot_index": 0,
      "terminal_state": "infra-failed",
      "variant": "P-min"
    },
    {
      "attempts": [
        {
          "attempt_index": 0,
          "backoff_ms": null,
          "charged": true,
          "ended_at": "2026-08-13T16:56:26.439880+00:00",
          "http_status": null,
          "outcome": "infra-failed",
          "reached_model": true,
          "reason": "mid-stream-death",
          "request_id": null,
          "started_at": "2026-08-13T16:56:20.140888+00:00"
        }
      ],
      "block_index": 1,
      "model_reaching_attempt_index": 0,
      "slot_index": 1,
      "terminal_state": "infra-failed",
      "variant": "P-min"
    },
    {
      "attempts": [
        {
          "attempt_index": 0,
          "backoff_ms": null,
          "charged": true,
          "ended_at": "2026-08-13T16:56:32.433179+00:00",
          "http_status": null,
          "outcome": "infra-failed",
          "reached_model": true,
          "reason": "mid-stream-death",
          "request_id": null,
          "started_at": "2026-08-13T16:56:26.440906+00:00"
        }
      ],
      "block_index": 1,
      "model_reaching_attempt_index": 0,
      "slot_index": 1,
      "terminal_state": "infra-failed",
      "variant": "P-q"
    }
  ],
  "retry_policy": {
    "max_unreachable_retries": 5,
    "rate_limit_backoff": {
      "max_attempts": 4,
      "max_total_ms": 60000
    }
  }
}
gpt-5.1 @low · cursor-xl — card · weather-card-benchmark