{
 "totals": {
  "metered_total_usd": 0.7652,
  "metered_rows": [
   "openai-gpt-6-astra",
   "kimi-k3"
  ],
  "calls_by_kind": {
   "arm": 43,
   "seat": 53,
   "probe": 1
  },
  "calls_total": 97,
  "registered_expected_metered_spend_usd": 0.703,
  "registered_ceiling_usd": 0.75,
  "over_the_registered_estimate": true,
  "over_the_registered_ceiling": true,
  "sentence": "the metered rows come to $0.7652 together, against a registered estimate of $0.70 and a ceiling of $0.75 — over both, by a cent and a half, and printed rather than rounded. No cap fired: the caps are $15.00 in all with $5.00 for the arms and $5.00 for the judges, and the count of cells any cap wrote a state word on is beside this sentence.",
  "what_a_total_is_not": "the metered total is the only total this bill draws. The subscription row's two estimates are never added into it — an estimate summed with a receipt is a receipt nobody holds — and the plan-included rows contribute $0.00* rather than $0.00, so there is nothing of theirs to add either."
 },
 "round": "two-frontiers-cove-r1",
 "kind": "bill",
 "schema": "s2s-bench-v1",
 "window": {
  "opened_utc": "2026-09-06T13:27:57Z",
  "closed_utc": "2026-09-06T17:03:52Z",
  "_note": "stamped UTC at BOTH ends, here and in the kit's index.json (the cold-scroll law).",
  "_closed_note": "stamped at the SEAL, per §10. The last call this round made was a judge sheet at 2026-09-06T15:50:29Z; nothing has been asked of any model since, and no lane after this one makes a call. The window therefore CONTAINS every call the round made, gate probes and the four discarded warmups included, which is why §10 stamps it at the commit rather than at the first call."
 },
 "records_window": {
  "earliest": "2026-09-06T13:30:36Z",
  "latest": "2026-09-06T14:00:27Z",
  "over": "every arm record this bill sums (the seats' own runs carry their windows in the carriage report)"
 },
 "cost_states": {
  "metered": "actual token counts priced at the CITED rate — an upper bound, and it says so",
  "plan-included": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
  "no-figure-held": "a subscription account holds no per-call receipt; the transport's own list-rate estimate prints, labelled as an estimate",
  "own-silicon": "our own hardware: no dollar exists for this row. An em dash, never a zero"
 },
 "plan_included_note": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
 "order_basis": "the REGISTERED roster order for the arms, then the registered panel order for the seats. Not a sort by any figure.",
 "rows": [
  {
   "id": "cli-claude-fable-5-1",
   "model": "claude-fable-5-1",
   "role": "reference arm",
   "cost_state": "no-figure-held",
   "cost_state_means": "a subscription account holds no per-call receipt; the transport's own list-rate estimate prints, labelled as an estimate",
   "transport_class": "agent-harness-cli",
   "calls": 12,
   "warmup_calls": 1,
   "records_by_leg": {
    "COVE": 9,
    "probe": 3
   },
   "tokens_in": 26156,
   "tokens_out": 5841,
   "reasoning_tokens": 2725,
   "window": {
    "earliest": "2026-09-06T13:30:36Z",
    "latest": "2026-09-06T13:50:41Z"
   },
   "registered_calls_cell": "9 arm + 1 effort + 3 quota = 13",
   "calls_basis": "the calls this bill sums, by leg, from the round's own append-only sent tree. §9's own calls cell for this row prints beside it: where the two differ, a probe registered at three moments ran at fewer, and the probe ledger names which.",
   "prompt_token_shape": "input + cache_creation + cache_read, summed the way this transport's OWN accounting sums it",
   "usd": 0.6461,
   "usd_unrounded": 0.6461115,
   "usd_registered_basis": 1.559,
   "registered_basis": {
    "usd_per_call": 0.12992,
    "basis_total_usd": 31.44,
    "basis_records": 242,
    "registered_sentence": "exhibit forty's $31.44 over 242 records = $0.12992/call",
    "read_from": "prereg/PREREG-COVE-LEG.md §9"
   },
   "shows_as": "$0.6461 by the transport's own per-call estimate, $1.5590 by §9's registered basis — neither is a receipt",
   "multiplication": "two estimates of a price nobody holds a receipt for, and both print: the transport's own per-call estimate of what these tokens would have cost on the API, summed over 12 calls = $0.6461; and §9's registered basis, 12 calls × $0.12992 = $1.5590 (exhibit forty's $31.44 over 242 records = $0.12992/call)",
   "kind": "arm"
  },
  {
   "id": "openai-gpt-6-astra",
   "model": "gpt-6-astra",
   "role": "reference arm",
   "cost_state": "metered",
   "cost_state_means": "actual token counts priced at the CITED rate — an upper bound, and it says so",
   "transport_class": "openai-api",
   "calls": 10,
   "warmup_calls": 1,
   "records_by_leg": {
    "COVE": 9,
    "probe": 1
   },
   "tokens_in": 15597,
   "tokens_out": 8140,
   "reasoning_tokens": 6666,
   "window": {
    "earliest": "2026-09-06T13:41:39Z",
    "latest": "2026-09-06T13:54:16Z"
   },
   "registered_calls_cell": "9 arm + 1 effort probe = 10",
   "calls_basis": "the calls this bill sums, by leg, from the round's own append-only sent tree. §9's own calls cell for this row prints beside it: where the two differ, a probe registered at three moments ran at fewer, and the probe ledger names which.",
   "prompt_token_shape": "the endpoint's own single input counter",
   "usd": 0.5183,
   "usd_unrounded": 0.518276,
   "usd_if_no_cache": 0.563,
   "cached_prompt_tokens": 4966,
   "shows_as": "$0.5183",
   "multiplication": "10,631 uncached in × $10.0/M + 4,966 cached in × $1.0/M + 8,140 out (incl. 6,666 reasoning) × $50.0/M = $0.5183. Had none of the input been cached the row would read $0.5630, which is what the round's own per-call ledger carries",
   "rates_usd_per_mtok": {
    "input": 10.0,
    "cached_input": 1.0,
    "output": 50.0
   },
   "pricing_read_date": "2026-09-05",
   "kind": "arm"
  },
  {
   "id": "cloud-glm-5-3",
   "model": "glm-5.3",
   "role": "reference arm",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "transport_class": "ollama-cloud",
   "calls": 11,
   "warmup_calls": 1,
   "records_by_leg": {
    "COVE": 9,
    "probe": 2
   },
   "tokens_in": 15719,
   "tokens_out": 29372,
   "reasoning_tokens": 0,
   "window": {
    "earliest": "2026-09-06T13:30:58Z",
    "latest": "2026-09-06T14:00:01Z"
   },
   "registered_calls_cell": "9 arm + 1 effort probe = 10",
   "calls_basis": "the calls this bill sums, by leg, from the round's own append-only sent tree. §9's own calls cell for this row prints beside it: where the two differ, a probe registered at three moments ran at fewer, and the probe ledger names which.",
   "prompt_token_shape": "the endpoint's own single input counter",
   "usd": 0.0,
   "usd_unrounded": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "kind": "arm"
  },
  {
   "id": "local-gemma4-26b",
   "model": "gemma4:26b",
   "role": "arm",
   "cost_state": "own-silicon",
   "cost_state_means": "our own hardware: no dollar exists for this row. An em dash, never a zero",
   "transport_class": "local-ollama",
   "calls": 10,
   "warmup_calls": 1,
   "records_by_leg": {
    "COVE": 9,
    "probe": 1
   },
   "tokens_in": 16070,
   "tokens_out": 1374,
   "reasoning_tokens": 0,
   "window": {
    "earliest": "2026-09-06T13:30:58Z",
    "latest": "2026-09-06T14:00:27Z"
   },
   "registered_calls_cell": "9 arm + 1 effort probe = 10",
   "calls_basis": "the calls this bill sums, by leg, from the round's own append-only sent tree. §9's own calls cell for this row prints beside it: where the two differ, a probe registered at three moments ran at fewer, and the probe ledger names which.",
   "prompt_token_shape": "the endpoint's own single input counter",
   "usd": null,
   "shows_as": "—",
   "multiplication": "none — our own hardware",
   "kind": "arm"
  },
  {
   "id": "gemma4-31b",
   "role": "judging seat",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "calls": 9,
   "tokens_in": 52615,
   "tokens_out": 7637,
   "read_from": [
    "results/cove/judge/two-frontiers-cove-r1/runs/RUN-gemma4-31b.json",
    "results/cove/judge/two-frontiers-cove-audition/AUDITION-gemma4-31b.json"
   ],
   "cap_usd": null,
   "usd": 0.0,
   "usd_unrounded": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "kind": "seat"
  },
  {
   "id": "mistral-large-3-675b",
   "role": "judging seat",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "calls": 9,
   "tokens_in": 51139,
   "tokens_out": 6816,
   "read_from": [
    "results/cove/judge/two-frontiers-cove-r1/runs/RUN-mistral-large-3-675b.json",
    "results/cove/judge/two-frontiers-cove-audition/AUDITION-mistral-large-3-675b.json"
   ],
   "cap_usd": null,
   "usd": 0.0,
   "usd_unrounded": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "kind": "seat"
  },
  {
   "id": "nemotron-3-ultra",
   "role": "judging seat",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "calls": 9,
   "tokens_in": 54257,
   "tokens_out": 6113,
   "read_from": [
    "results/cove/judge/two-frontiers-cove-r1/runs/RUN-nemotron-3-ultra.json",
    "results/cove/judge/two-frontiers-cove-audition/AUDITION-nemotron-3-ultra.json"
   ],
   "cap_usd": null,
   "usd": 0.0,
   "usd_unrounded": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "kind": "seat"
  },
  {
   "id": "kimi-k3",
   "role": "judging seat",
   "cost_state": "metered",
   "cost_state_means": "actual token counts priced at the CITED rate — an upper bound, and it says so",
   "calls": 9,
   "tokens_in": 49139,
   "tokens_out": 6632,
   "read_from": [
    "results/cove/judge/two-frontiers-cove-r1/runs/RUN-kimi-k3.json",
    "results/cove/judge/two-frontiers-cove-audition/AUDITION-kimi-k3.json"
   ],
   "cap_usd": 5.0,
   "usd": 0.2469,
   "usd_unrounded": 0.24689999999999998,
   "shows_as": "$0.2469",
   "multiplication": "49,139 in × $3.0/M + 6,632 out × $15.0/M = $0.2469",
   "rates_usd_per_mtok": {
    "input": 3.0,
    "cached_input": 0.3,
    "output": 15.0
   },
   "kind": "seat"
  },
  {
   "id": "deepseek-v4-pro",
   "role": "judging seat",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "calls": 8,
   "tokens_in": 42893,
   "tokens_out": 6126,
   "read_from": [
    "results/cove/judge/two-frontiers-cove-r1/runs/RUN-deepseek-v4-pro.json",
    "results/cove/judge/two-frontiers-cove-audition/AUDITION-deepseek-v4-pro.json"
   ],
   "cap_usd": null,
   "usd": 0.0,
   "usd_unrounded": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "kind": "seat"
  },
  {
   "id": "glm-5.3",
   "role": "judging seat — NOT-CARRIED at its audition, retired for the round",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "calls": 2,
   "tokens_in": 7999,
   "tokens_out": 3917,
   "read_from": [
    "results/cove/judge/two-frontiers-cove-audition/AUDITION-glm-5.3.json"
   ],
   "cap_usd": null,
   "usd": 0.0,
   "usd_unrounded": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "kind": "seat"
  },
  {
   "id": "qwen3.5-397b",
   "role": "judging seat",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "calls": 7,
   "tokens_in": 40288,
   "tokens_out": 5595,
   "read_from": [
    "results/cove/judge/two-frontiers-cove-r1/runs/RUN-qwen3.5-397b.json",
    "results/cove/judge/two-frontiers-cove-audition/AUDITION-qwen3.5-397b.json"
   ],
   "cap_usd": null,
   "usd": 0.0,
   "usd_unrounded": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "kind": "seat"
  },
  {
   "id": "mistral-large-3-675b",
   "kind": "probe",
   "model": "mistral-large-3:675b",
   "role": "the outside read (G-OUTSIDE-READ)",
   "cost_state": "plan-included",
   "cost_state_means": "$0.00* — no marginal charge, not free: a flat monthly plan whose price is the account's and not this round's. The asterisk is load-bearing.",
   "calls": 1,
   "tokens_in": 41570,
   "tokens_out": 3735,
   "usd": 0.0,
   "shows_as": "$0.00*",
   "multiplication": "none — plan-included",
   "read_from": [
    "prereg/receipts-cove/g-outside-read-cove.json"
   ],
   "verdict_as_filed": "PASS"
  }
 ],
 "metered_total_usd": 0.7652,
 "metered_rows": [
  "openai-gpt-6-astra",
  "kimi-k3"
 ],
 "caps": {
  "registered_usd": {
   "arms": 5.0,
   "judges": 5.0,
   "probes": 2.0,
   "reserve": 3.0,
   "total": 15.0
  },
  "seat_caps_usd": {
   "kimi-k3": 5.0
  },
  "arm_caps_usd": {
   "openai-gpt-6-astra": 15.0
  },
  "metered_total_usd": 0.7652,
  "cap_cells_written": 0,
  "sentence": "every dollar above is an UPPER BOUND and says so; a cap firing writes a state word on every remaining cell, never an estimate and never a zero. The count of cells carrying that state word is beside this sentence, so 'no cap fired' is a comparison rather than a claim."
 },
 "balances": "prereg/receipts-cove/balances-cove.json — an operator's own reading of the two metered balances, re-read at registration",
 "integers": {
  "arms": 4,
  "seats": 7
 }
}