Touchstone Assay
Dated SIU — SIU-2026a
$0.0169
21 Sep 2026 provisional
Correction notice — published after this print, disclosing a fact that does not change dated_siu:

3 corrections: 1 precision, 2 pricing

  • 25 Sep 2026 — dated_siu previously rounded to a fixed 4 decimal places rather than 4 significant figures — found live and fixed the same day, 2026-09-25 (docs/methodology.md's Rounding section). At this print's price level that was too coarse to show real day-to-day movement. This print's signed body is never edited — the published figure (0.0169) stands. Recomputed from this print's own published basket_costs and weights (Σ weight × basket_cost, rounded under the corrected 4-significant-figure rule, reproducible by anyone from this print's own published figures alone), the real value is 0.01692.
  • 25 Sep 2026 — This print was published under the belief that cached-input pricing (rules-2026-09-13, commit 29d4be2) was already live — docs/methodology.md stated this as fact. It was not: buildModelInputs never forwarded a price snapshot's price_cached_in_usd_per_1m to the cost formula, for any model, from any source, until this was found and fixed live 2026-09-25 (commit 327d7b9, docs/methodology.md's rules-2026-09-25b). For this print specifically, real nonzero cached_input usage on gemini-3.1-pro-preview, grok-4.6 — already present in this print's own run records — was still priced at zero in this print's own signed cost_of_production_usd and basket_costs, contrary to what the methodology this print's date falls under claimed. This print's signed body is never edited. The exact, per-model corrected dollar figure is not computed here — that quantified backfill is separate, deferred follow-up work; this note discloses the structural fact, not yet the number.
  • 26 Sep 2026 — Quantified follow-up to this print's own cached-input-pricing correction note above (the "separate, deferred follow-up work" that note left open). Corrected figures, computed from this print's own real run records and price snapshot with the fix applied (commit 327d7b9), not from new data: this print's dated_siu was originally reported as $0.0169; the corrected figure is $0.0171. gemini-3.1-pro-preview's own basket cost was originally reported as $0.038780; the corrected figure is $0.039762. grok-4.6's own basket cost was originally reported as $0.017182; the corrected figure is $0.017412. This print itself is not recomputed or replaced — the original stands, per the revision policy's unconditional never-edit guarantee. These are also now the exact figures packages/print's own verify command (recomputing under this print's own historical rounding rule, and from the same run records the blended Dated SIU print for this date uses, filtered to this series' own registered models) reports.
▼ -0.6% since 20 Sep 2026 — Cheaper than the previous print. Falling prices are the expected direction for this index over time.
Equal weighting, disclosed: basket costs span 7.3× across the 6 qualifying constituents. 3 of 6 constituents (50%) sit at or above the mean basket cost and together account for 78% of the summed basket cost Dated SIU is computed from — equal weighting still gives each of them the same 1÷6 nominal share. Dated SIU therefore tracks the more expensive end of the basket more closely than a demand-weighted index would. See the methodology's Weighting section, and Frontier SIU / Commodity SIU, the tier prints published alongside this one for exactly this reason.

Exchange-rate table

ModelUSD / SIUSpread to indexSIU per $1
claude-sonnet-5 $0.0234 +38% 42.7
gpt-5.1 $0.0092 -46% 108.9
gemini-3.1-pro-preview $0.0388 +129% 25.7
grok-4.6 $0.0172 +2% 58.2
claude-haiku-4-5 $0.0077 -55% 130.5
gpt-5.4-mini $0.0053 -68% 187.0

Floor & market spread

Not yet published — the floor needs a measured hardware-rental rate for the reference configuration, which this print does not yet include.

Sensitivity

The print recomputed under alternative cache and batch-discount policies, so the effect of those policy choices is stated rather than hidden inside one number.

Policy variantDated SIUDeltaApplies to
cache-enabled: 40% of T2 input at 10% of input price $0.0122 -27.8% claude-sonnet-5, gpt-5.1, gemini-3.1-pro-preview, grok-4.6, claude-haiku-4-5, gpt-5.4-mini
batch-discount: 50% off input and output $0.0085 -50.0% claude-sonnet-5, gpt-5.1, gemini-3.1-pro-preview, grok-4.6, claude-haiku-4-5, gpt-5.4-mini

Methodology, signature & anchor

Methodology
v0-draft
Cost of production
$1.6199
Signature
0x54c5afb6…1c9f4823
Public key
0x030c598e…e25cb299
Anchor tx — Base Sepolia
0xd24be340…1f800eb9 (base-sepolia)
Anchor tx — Arc Testnet
not anchored: anchoring on Arc Testnet began with the print dated 2026-10-15
Raw runs
https://github.com/petrkrulis2022/dated-siu-inference/tree/main/data/runs/2026-09-21-frontier

Dated SIU is an assessment of qualifying inference work under this basket version, model/provider configuration and quality gate — not a prediction of production agent costs, which vary with workflow design, tool reliability, retry policy and context handling.

Verify this yourself

Don't take this page's word for it — every value below is checkable independently. Read the steps, run one command, or have an agent read verify.json.

The body hash

Body hash
0xddd0fbefcb18038e0254e38cbe041172b415b7898b257c8696ade93b404ce5c9
Recovered signer
0x284ff2f8605ff8afeda6959b856bb7e6d48f845a

The body hash is made like this:

  1. Download this print's JSON: https://raw.githubusercontent.com/petrkrulis2022/dated-siu-inference/main/data/prints/2026-09-21-frontier.json.
  2. Remove exactly these five fields and keep everything else as published: signature, public_key, anchor, superseded_by, correction_notes. The last three are added after signing, which is why they cannot be part of the hash.
  3. Canonicalise what is left with RFC 8785 (JCS): keys sorted, no whitespace, numbers and strings as JSON writes them.
  4. Take the keccak256 hash of the UTF-8 bytes of that text (the original Keccak padding Ethereum uses, not NIST SHA3-256). Written as 0x and 64 hex digits, it equals the body hash above.
Or run one command 
curl -O https://prints.touchstoneassay.com/verify-print.mjs
node verify-print.mjs 2026-09-21-frontier

Plain Node 18 or later, nothing to install, no keys. It does the four steps above, then looks the anchor transaction up on the chain's explorer and looks for a PrintPosted event carrying your hash. Exit code 0 means it matched. It does not check the signature: see the recovered signer above and the publisher below.

Anchored on Base Sepolia

TouchstoneAttestation
0xF60701793eD168ffd6e818e1DCcb600393297190 (base-sepolia · chain id 84532)
Publisher (fixed in the contract)
0x284ff2F8605Ff8AFeDa6959B856Bb7E6d48f845a — the recovered signer above is this publisher
Anchor tx
0xd24be340228bea3a258f9949116cf84f29e3afc4c11802d07cb00ee01f800eb9 (base-sepolia)
  1. Open the contract link above on its explorer, go to "Read Contract", call publisher(), and confirm it equals the recovered signer shown above.
  2. On the same "Read Contract" tab, call postedAt(bytes32) with the body hash above and confirm it returns a non-zero timestamp: the Unix time, in seconds, at which this hash was first recorded.
  3. Open the anchor transaction, go to its Logs tab, and confirm a PrintPosted event whose bodyHash equals the body hash above.

Anchored on Arc Testnet

TouchstoneAttestation
0x12b886b043feABc3d90bBae3ae206d22b208160d (arc-testnet · chain id 5042002)
Publisher (fixed in the contract)
0x284ff2F8605Ff8AFeDa6959B856Bb7E6d48f845a — the recovered signer above is this publisher
Anchor tx
not anchored: anchoring on Arc Testnet began with the print dated 2026-10-15

All prints