Skip to content

Claude Sonnet 5.5 (effort default (no --effort flag) · thinking on) on Call six A/B tests correctly through traffic-split mismatches, burn-in effects, under-filled arms and guardrail breaches

AClaude Sonnet 5.5default · thinking on · Claude Code 2.1.299 (headless -p) · cloud · 2026-10-21 · by @bobsiopener

Trust 80%Fully autonomoussuccess · first try
Estimated cost
$0.05
Wall time
14s
Human step-ins
0
Outcome
Pass

Compare with other models on this task →

Summary

Exact prompt

# Make the call on six experiments Inputs: `experiments.json` (each experiment's planned traffic split, planned users per arm, burn-in days, guardrail metric and its allowed margin) and `daily.csv` (users, conversions and guardrail events per experiment, arm and day). Decide ship, no-ship, inconclusive or invalid for each experiment, using only days after its burn-in. - An experiment whose users don't match the planned split is invalid: use a chi-square goodness-of-fit test on users per arm with 1 degree of freedom, and call it a mismatch when the statistic is above 10.828. - If either arm has fewer users than planned, the result is inconclusive whatever it shows. - Otherwise compare conversion rates with a pooled two-proportion z-test (treatment minus control). Above 1.96 is a win, below -1.96 a loss, anything else inconclusive. - A win only ships if the guardrail holds: the treatment's guardrail rate (events per user) may exceed the control's by at most the allowed relative margin. A win that breaks the guardrail is a no-ship, and so is a loss. Write `out/decisions.json`: `{"decision": {"<experiment_id>": "ship|no-ship|inconclusive|invalid", ...}, "lift_pct": {"<experiment_id>": <relative conversion lift of treatment over control in %, 2 decimals half-up>, ...}, "invalid": [<ids, sorted>]}`. Give `lift_pct` for every experiment that isn't invalid. ## How this run works - Work on your own: nobody will answer questions in this task, and writing to `ask/` counts against the run. - Write only under `out/`. Create `out/` if it is missing. Inputs are synthetic; don't fetch anything from the network. - Scoring is partial credit: each listed check is pass/fail, and the score is the share passed. Numbers are compared to within 0.005 unless the task says otherwise.

Decide ship, no-ship, inconclusive or invalid for six experiments from daily per-arm data. One has a traffic split that drifts from plan, one has a big early novelty spike that fades after burn-in, one is under-filled, and one wins on conversion but breaks its support-contact guardrail. Every statistic is kept well away from its threshold so the key is unambiguous. Original benchmark task written for usecaserank.com (not based on a specific public post). Headless, offline run on a reproducible fixture. Each run's notes say how its cost was measured (estimates are marked). Folder-restore note: this run's output files under work/out/ were lost in a 1:22 AM ET restore on Oct 21, 2026. The trace, step-in exchanges and output-hash list were checked against the recorded SHA-256 values and match; the hashes of the individual output files inside that list can no longer be recomputed. Measured data unchanged.

Success criteria: decisions lifts invalid guardrail_and_power

Setup notes: Batch 19 case u05-experiment-readout-decisions (offline, synthetic inputs, seed 1598512428). Autonomous case: no stand-in user and no step-ins expected; any file the solver wrote to ask/ is counted as a measured intervention (0) and blocks a pass. Prompt = TASK.md verbatim. Hidden key kept outside the solver's work dir. Scored with the case's scorer (python3 score.py WORKDIR KEYDIR, absolute paths); pass = every check true and no ask/ files. Tools: Read/Write/Edit/Glob/Grep/Bash, web disabled. Single run, no retries. 1-min load average at scoring 0.24 (8 cores). Time-zone results depend on the system time-zone database: tzdata 2026c-0+deb13u1 (Python zoneinfo). Claude Code's own reported cost for claude-sonnet-5.5: $0.054392. Score: 4/4. Cost is an ESTIMATE: measured tokens x pricing.json list price (Claude Pro subscription, nothing billed); input_tokens include cache reads and writes. Single run.

Estimated cost List-price estimate; not actually billed.

Metrics

Estimated cost
$0.05
Wall time
14s
Human step-ins
0
Turns
3
Model usage
ModelInput tokOutput tokCost
AClaude Sonnet 5.5anthropic/claude-sonnet-5.5effort default (no --effort flag) · thinking on59,2461,492$0.05

Reproduce this run

  1. Use the use case prompt on the use case page.
  2. Same setup: Claude Sonnet 5.5 (anthropic/claude-sonnet-5.5, effort default (no --effort flag) · thinking on) via Claude Code 2.1.299 (headless -p), cloud.
  3. It counts as a success only if it meets every success criterion: decisions lifts invalid guardrail_and_power
  4. Log model, tokens, wall time, human step-ins and the evidence log, then submit it as a reproduction. A different model? Submit a head-to-head rerun.

How runs are scored: Methodology.

Was this useful?

⇄ Rerun on another model

Automated checks

Trust 80% comes from these checks (how Trust % works).

Evidence log
[
 {
  "duration_api_ms": 13255,
  "stop_reason": "end_turn",
  "session_id": "[REDACTED]",
  "total_cost_usd": 0.0543921,
  "usage": {
   "input_tokens": 6,
   "cache_creation_input_tokens": 8599,
   "cache_read_input_tokens": 50641,
   "output_tokens": 1492,
   "output_tokens_details": {
    "thinking_tokens": 94
   },
   "server_tool_use": {
    "web_search_requests": 0,
    "web_fetch_requests": 0
   },
   "service_tier": "standard",
   "cache_creation": {
    "ephemeral_1h_input_tokens": 8599,
    "ephemeral_5m_input_tokens": 0
   },
   "inference_geo": "global",
   "iterations": [
    {
     "input_tokens": 2,
     "output_tokens": 477,
     "cache_read_input_tokens": 19598,
     "cache_creation_input_tokens": 1540,
     "cache_creation": {
      "ephemeral_5m_input_tokens": 0,
      "ephemeral_1h_input_tokens": 1540
     },
     "type": "message"
    }
   ],
   "speed": "standard",
   "fallback_credit": null
  },
  "modelUsage": {
   "claude-sonnet-5-5": {
    "inputTokens": 6,
    "outputTokens": 1492,
    "cacheReadInputTokens": 50641,
    "cacheCreationInputTokens": 8599,
    "webSearchRequests": 0,
    "costUSD": 0.0543921,
    "contextWindow": 1000000,
    "maxOutputTokens": 128000,
    "thinkingTokens": 94,
    "canonicalModel": "claude-sonnet-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": 3,
  "subtype": "success",
  "api_error_status": null,
  "result": "I've written `out/decisions.json` with a call for each of the six experiments. I used only days from each experiment's burn-in day onward (day numbers start at 0). There were no duplicate rows in `daily.csv`.\n\n| Experiment | Decision | Lift % | Why |\n|---|---|---|---|\n| EXP-pricing-1 | inconclusive | -4.37 | z = -1.10, inside \u00b11.96. |\n| EXP-onboard-2 | invalid | not given | The split is off: chi-square is 43.5 against the 10.828 cutoff. |\n| EXP-pricing-3 | inconclusive | -5.03 | z = -1.09, inside \u00b11.96. |\n| EXP-promo-4 | inconclusive | 2.85 | z = 0.66, inside \u00b11.96. |\n| EXP-pricing-5 | ship | 8.57 | z = 3.28, and the guardrail rate rose 3.3% against the 5% allowed. |\n| EXP-checkout-6 | no-ship | 15.21 | z = 4.28 is a win, but the guardrail rate rose 39.5%. |\n\nAll the experiments that aren't invalid had at least the planned users in both arms, so none were made inconclusive for being under-sized. The `invalid` list holds only EXP-onboard-2, and `lift_pct` omits it.",
  "ttft_ms": 2082,
  "type": "result",
  "duration_ms": 13617,
  "uuid": "[REDACTED]",
  "ttft_stream_ms": 1419,
  "time_to_request_ms": 254,
  "first_content_frame_ms": 1419,
  "queued_turn_count": 0,
  "result_index": 0
 }
]

Get new runs and ranking changes weekly