ScaffoldBench

RunSonnet 5.5repeat 1

Propose-select on Redirect the order that has not shipped

claude-sonnet-5-5__propose-select__ecommerce-easy__r1

Score

1.000

passed
design ran as designed
Cost
$0.129
Latency
27.5s
setup 0.0s, grading 0.0s
Turns
24
Tool calls
18
Subagents
3
Tokens in / out
65k / 4.6k
39,988 read from cache

Did the design run?

every check held
  • Run completed: heldharness.run-completedThe environment was built, the architecture returned and grading finished, none of them throwing.
  • No API failures: heldharness.apiNo model call failed at the API.
  • Every call ended: heldharness.calls-endedEvery call ended with a result, or was stopped by the run's own limits.
  • No harness tool denied: heldharness.harness-toolsThe sandbox guard never denied a harness tool, nor a tool the session itself offered.
  • Tool servers connected: heldharness.mcp-connectedEvery tool server the harness gave a call was connected when the session started.
  • No foreign tool servers: heldharness.mcp-hermeticNo call loaded a tool server the harness did not give it.
  • Builtin tools: heldharness.builtin-toolsA check recorded with the run.
  • Structured replies arrived: heldharness.structured-outputEvery call that asked for structured output and ran to the end got it.
  • No silent fallbacks: heldharness.not-degradedNo phase fell back on a default (an empty plan, a default winner) instead of its output.
  • All three proposers ran: heldpropose-select.proposalsThree proposers ran in parallel, each from its own angle, and none was starved while another ran.
  • The selector chose: heldpropose-select.selectionThe selector picked a proposal by a valid number, read from its schema reply and not a default.
  • Each call was shown the right tools: heldpropose-select.toolsEach proposer was shown the read-only tools, the selector none, and the executor the full toolset.

Phases

Proposal 1 (policy-first): 0.0s to 11.7s, finished, 4 tool calls, $0.020
Proposal 2 (state-first): 0.0s to 12.1s, finished, 4 tool calls, $0.027
Proposal 3 (outcome-first): 0.0s to 9.9s, finished, 4 tool calls, $0.027
Select: 12.1s to 19.1s, finished, 0 tool calls, $0.026
Execute: 19.1s to 27.5s, finished, 6 tool calls, $0.029

Drawn from the phase records: timing and cost per call, no tool calls or messages.

PhaseDurationCostTurnsToolsModel
Proposal 1 (policy-first)11.7s$0.02054claude-sonnet-5-5
Proposal 2 (state-first)12.1s$0.02754claude-sonnet-5-5
Proposal 3 (outcome-first)9.9s$0.02754claude-sonnet-5-5
Select7.0s$0.02620claude-sonnet-5-5
Execute8.4s$0.02976claude-sonnet-5-5

The call-by-call replay is not published for this run. Traces are large, so the repository keeps 9 of this run set’s 900 and leaves the rest out; the score, phases and final answer on this page come from the committed run record.

Runs you can replay

Record this cell yourself: npm run bench -- run --strategies propose-select --tasks ecommerce-easy --models claude-sonnet-5-5 --repeats 1

Score breakdown

  • passThe order that has not shipped carries the new address
    1.00weight 0.43

    O-1001 address: 7 Quarry Rise, Sheffield S11 8XZ, UK

  • passThe edit did not spill onto the order already in transit
    1.00weight 0.29
  • passLooked up which order was still open before editing one
    1.00weight 0.14
  • passChecked the state of the order already in transit
    1.00weight 0.14

Penalties

No penalties were applied.

Scorer metadata

tool calls
18
failed tool calls
0
distinct tools
4

Final answer

I changed the shipping address on O-1001, the order that hasn't shipped, and left the in-transit order O-1002 as it was.

  • O-1001: The address was 12 Alder Lane, Bristol BS1 4TR, UK. It is now 7 Quarry Rise, Sheffield S11 8XZ, UK. I re-read the order afterwards. It is still pending, both lines are still pending, and the payment of 16,700 cents and the refund total of 0 are unchanged.
  • O-1002: I did not touch it. It is still shipped, with line L1 shipped on 2026-07-07, and still shows 88 Old Mill Road, Bristol BS3 2QQ, UK. The handbook says the address of a shipped order is never changed, and a change couldn't redirect the parcel anyway.

I made no cancellations, refunds, restocks or store-credit changes.

Provenance of the numbers on this page

run set
scaffold-or-scale-v1
models
Haiku 4.5, Sonnet 5.5, Opus 5.5
effort
high
repeats
2
fixtures
replay
runs
900 / 900
strategy cost
$354.44
judge cost
$9.19
wall clock
9h 48m
git
fce5807b2d

Cost excludes judge calls, which are tracked separately so a scoring model never flatters or penalises an architecture. Every figure on this site is read from results/ at build time.