ScaffoldBench

RunSonnet 5.5repeat 1

Propose-select on Rebook onto the earliest available flight and settle the difference

claude-sonnet-5-5__propose-select__airline-medium__r1

Score

1.000

passed
design ran as designed
Cost
$0.184
Latency
32.1s
setup 0.0s, grading 0.0s
Turns
31
Tool calls
25
Subagents
3
Tokens in / out
84k / 6.5k
49,631 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 13.4s, finished, 6 tool calls, $0.039
Proposal 2 (state-first): 0.0s to 13.1s, finished, 6 tool calls, $0.038
Proposal 3 (outcome-first): 0.0s to 13.5s, finished, 6 tool calls, $0.038
Select: 13.5s to 22.9s, finished, 0 tool calls, $0.033
Execute: 22.9s to 32.1s, finished, 7 tool calls, $0.036

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

PhaseDurationCostTurnsToolsModel
Proposal 1 (policy-first)13.4s$0.03976claude-sonnet-5-5
Proposal 2 (state-first)13.1s$0.03876claude-sonnet-5-5
Proposal 3 (outcome-first)13.5s$0.03876claude-sonnet-5-5
Select9.4s$0.03320claude-sonnet-5-5
Execute9.2s$0.03687claude-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 airline-medium --models claude-sonnet-5-5 --repeats 1

Score breakdown

  • passReservation moved to the earliest flight with an available seat, same cabin
    1.00weight 0.43

    booked on AA204, expected AA204

  • passExactly one fee, the fare difference for the flight actually booked
    1.00weight 0.43

    fare difference owed on the booked flight: 8000 cents

  • passExactly one economy seat moved from the old flight to the new one
    1.00weight 0.14

Penalties

No penalties were applied.

Scorer metadata

tool calls
25
failed tool calls
0
distinct tools
8

Final answer

Dana Whitfield's reservation R2087 is now on AA204, departing SFO at 09:05 on 2026-08-03, still in economy. She owes $80.00 for the change, and that is now charged.

  • Why AA204: It is the earliest economy departure with a seat. UA88 (06:15) has no economy seats left, and AA918 (07:30) is cancelled. AA204 had 4 economy seats left when I rechecked.
  • Fare difference: The AA204 economy fare is $390.00 and she paid $310.00. I recorded the $80.00 difference as one fare_difference ledger entry (L1).
  • Change fee: The $75.00 fee is waived because Dana is a gold member, so I charged none.
  • Bags: Her one checked bag is unchanged, so there is no baggage fee.
  • Other records: I didn't touch any other booking, passenger record or flight.

The system date is 2026-10-05, which is after the 2026-08-03 travel date. The booking tool accepted the change anyway and nothing in the policy blocked it, so I went ahead. You may want to check that date.

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.