Versioned benchmark collection

Robot pick-and-place

Public ledger · prompt hash recorded · 5 measured runs across 1 target. Each entry is a single recorded attempt; repeat-run reliability is not established.

Robot pick-and-place public evidence
Robot pick-and-place public evidence
5Measured runs
1Frozen targets
5Model families
5Configurations
Collection scope

Recorded evidence, without a synthetic score.

Runs share the same frozen target and prompt within each test, with disclosed provider-specific stacks. Costs are recorded estimates as labelled in each ledger, not subscription invoices.

Targets

  • Apple Carry

Model coverage

  • GPT-6 Astra · Max
  • GPT-6 Sol · Max
  • Grok 4.7 · xhigh
  • MiMo V2.6 Pro · Highest
  • Claude Opus 5.5 · Max

Evidence available

  • 5 runs on prompt 699ebc2b02b4
  • Prompt SHA-256 699ebc2b02b4dc3f8e22b498948bd2af7ad7e94b10b288f7629589fe875e092b
  • Exact prompt file path was not published

Still missing

  • No released Builder project, harness package, or skill download.
  • Exact prompt file path was not published with this test.
  • 2 related records are listed separately and are not attached to this prompt.
  • Repeat-run reliability evidence is not established.
Public receipts

5 measured run ledgers

Open any row for its recorded workflow cost, timing, token usage, stack disclosure, artifact integrity, and explicit evidence gaps.

Prompt provenance

Exact prompt receipts

A path is shown only when the published record includes the prompt hash that identifies it. Campaign evidence and projections keep the hash gap visible.

apple-carry

SHA-256 699ebc2b02b4dc3f8e22b498948bd2af7ad7e94b10b288f7629589fe875e092b

Exact prompt file path was not published.

Separate records

Related records with unknown prompt provenance

These published records stay separate from the prompt receipts above. Unknown provenance is why they are listed here.