{ "schema": "secondrun.campaign-intake.v1", "generated_at": "2026-09-24T06:12:57.413223+00:00", "status": "STAGED_REAL_EVIDENCE", "challenge_status": "HOLD", "challenge_hold_reasons": [ "No ClusterMAX-specific frozen prediction plan or medal binding supplied.", "Native campaign acceptance rules are preserved, not retroactively changed to the challenge endpoint.", "These one-GPU services do not establish a managed-cluster comparison.", "Provider coverage below the eight-provider study floor.", "Whole-allocation invoice reconciliation is incomplete; no modeled price substituted." ], "source_commit": "cbc7e234fbb39fbde2fdec93bba42141d899c206", "scored_runs": 3, "providers": [ "digitalocean", "hotaisle" ], "distinct_scored_allocations": 2, "runs": [ { "run_id": "run3-scored-a-t0-20260924", "role": "scored_arm", "provider": "hotaisle", "allocation_id": "hotaisle-enc1-gpuvm004", "service_id": "vm-mi300x-1x", "region": "enc1", "gpus": 1, "model": { "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8", "revision": "dcaee4d4dfc5ee71ad501f01f530e5652438fde0", "precision": "FP8", "bytes": 31187041238 }, "runtime": { "image": "vllm/vllm-openai-rocm:v0.30.0", "digest": "sha256:2e7da1ad1c66836802072588adea75f9f4991da5f9545b4318e91d422c22ce6a", "engine": "vllm", "engine_version": "0.30.0", "attention_backend": [ "ROCM_ATTN" ], "linear_kernel": [ "AiterFp8BlockScaledMMKernel" ] }, "workload_id": "sha256:6c508237082261ac927b96825e2f097782389b1cebc3f78360f0665a8db3501f", "counts": { "attempted": 8622, "completed": 8622, "correct": 4371, "accepted": 4336 }, "request_slots": 8622, "p95_ttft_including_queue_ms": 170.97414731979364, "p99_ttft_including_queue_ms": 572.6243519782902, "p95_e2e_including_queue_ms": 12362.79807090759, "acceptance_rule": { "basis": "run3/PREREG.md: accepted = completed AND EvalPlus base+plus pass AND first text <= 1 s AND end <= 60 s from the SCHEDULED arrival (queue_time + ttft, queue_time + latency in detailed.json)", "ttft_ms": 1000, "e2e_ms": 60000, "correctness": true, "queue": true }, "evaluator_criterion": "evalplus-0.3.1-human-v0.1.10-mbpp-v0.2.0-base-and-plus-raw-completion-v1", "clocks": { "t_request": null, "t_ssh": "2026-09-24T00:10:57Z", "t_ready": "2026-09-24T00:45:07Z", "t_work_start": "2026-09-24T00:45:11Z", "t_work_end": "2026-09-24T01:45:19Z", "t_released": null, "sources": { "t_request": { "source": "ledger-times.json", "precision": "operator-recorded" }, "t_ssh": { "source": "arm.py --t-ssh (observed before the script started)", "precision": "1 s" }, "t_ready": { "source": "ledger-times.json: /v1/models listed the model", "precision": "1 s" }, "t_work_start": { "source": "ledger-times.json: replay child started", "precision": "1 s" }, "t_work_end": { "source": "ledger-times.json: replay child exited (or the finally-block stamp after an interrupt)", "precision": "1 s" }, "t_released": { "source": "closure.json (provider confirmation; container stop is not release)", "precision": "operator-recorded" }, "t_script_start": "2026-09-24T00:42:43Z", "t_script_end": "2026-09-24T01:45:22Z" }, "null_reasons": { "t_request": "not in closure.json: arm.py cannot observe the capacity request; the operator must supply it", "t_released": "not in closure.json: provider release is external to the seat (arm.py note: container stop is not release)" } }, "native_money": { "currency": "USD", "list_rate_per_gpu_hr": 2.99, "billing": "per-minute", "payer": "hotaisle-credit", "modeled_minutes": null, "modeled_minutes_basis": null, "modeled_usd": null, "modeled_lower_bound_minutes": 94.37, "modeled_lower_bound_usd": 4.7026, "modeled_lower_bound_basis": "t_ssh -> t_work_end at list; release and/or request unknown, so the billable window is at least this long", "billed_usd": null, "billed_ref": null, "credits_usd": null, "energy": { "watts_mean": null, "kwh": null, "tariff_usd_per_kwh": null, "usd": null, "meter": null, "reason": "cloud seat; no meter" }, "null_reasons": { "modeled_minutes": "t_request and/or t_released unknown (closure.json)", "modeled_minutes_basis": "see modeled_minutes", "modeled_usd": "buyer's window unknown; see modeled_lower_bound_usd", "billed_usd": "invoice not posted / not entered in closure.json", "billed_ref": "see billed_usd", "credits_usd": "not entered in closure.json (PREREG: credit, modeled and invoice are three fields)" } }, "evidence": { "ledger_sha256": "df975644f6c812e9d02cc157a970b190e9630271ded87070a61a619560a422ca", "detailed_sha256": "ef907e4cbe45832890109c4ca0d37851ad8e4f087beb70421506af897e6a67ab", "evaluation_sha256": "f5e2fe316996ac26d953fb5ebe6d0415b2debd9664c1bc74a309776748ff08ca", "mapping_sha256": "7ada98a8ee5cfea39b93a85827785a54fed20fd37f3a04770c4d699bac9d4b95", "ledger_path": "campaign/results/run3-scored-a-t0/ledger/run3-run3-scored-a-t0.ledger.json", "directory": "campaign/results/run3-scored-a-t0" }, "evidence_scope": "Exact supplied files and judgment vector checked; outputs were not re-executed or independently graded.", "billing_status": "AWAITING_ALLOCATION_BILL", "whole_bill_share_usd": null }, { "run_id": "run3-scored-a-t1-20260924", "role": "scored_arm", "provider": "hotaisle", "allocation_id": "hotaisle-enc1-gpuvm004", "service_id": "vm-mi300x-1x", "region": "enc1", "gpus": 1, "model": { "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8", "revision": "dcaee4d4dfc5ee71ad501f01f530e5652438fde0", "precision": "FP8", "bytes": 31187041238 }, "runtime": { "image": "vllm/vllm-openai-rocm:v0.30.0", "digest": "sha256:2e7da1ad1c66836802072588adea75f9f4991da5f9545b4318e91d422c22ce6a", "engine": "vllm", "engine_version": "0.30.0", "attention_backend": [ "ROCM_AITER_FA" ], "linear_kernel": [ "AiterFp8BlockScaledMMKernel" ] }, "workload_id": "sha256:6c508237082261ac927b96825e2f097782389b1cebc3f78360f0665a8db3501f", "counts": { "attempted": 8622, "completed": 8615, "correct": 4334, "accepted": 4292 }, "request_slots": 8622, "p95_ttft_including_queue_ms": 216.9149875640859, "p99_ttft_including_queue_ms": 1123.4679889678966, "p95_e2e_including_queue_ms": 12809.249830245972, "acceptance_rule": { "basis": "run3/PREREG.md: accepted = completed AND EvalPlus base+plus pass AND first text <= 1 s AND end <= 60 s from the SCHEDULED arrival (queue_time + ttft, queue_time + latency in detailed.json)", "ttft_ms": 1000, "e2e_ms": 60000, "correctness": true, "queue": true }, "evaluator_criterion": "evalplus-0.3.1-human-v0.1.10-mbpp-v0.2.0-base-and-plus-raw-completion-v1", "clocks": { "t_request": null, "t_ssh": "2026-09-24T00:10:57Z", "t_ready": "2026-09-24T01:49:02Z", "t_work_start": "2026-09-24T01:49:07Z", "t_work_end": "2026-09-24T02:49:15Z", "t_released": null, "sources": { "t_request": { "source": "ledger-times.json", "precision": "operator-recorded" }, "t_ssh": { "source": "arm.py --t-ssh (observed before the script started)", "precision": "1 s" }, "t_ready": { "source": "ledger-times.json: /v1/models listed the model", "precision": "1 s" }, "t_work_start": { "source": "ledger-times.json: replay child started", "precision": "1 s" }, "t_work_end": { "source": "ledger-times.json: replay child exited (or the finally-block stamp after an interrupt)", "precision": "1 s" }, "t_released": { "source": "closure.json (provider confirmation; container stop is not release)", "precision": "operator-recorded" }, "t_script_start": "2026-09-24T01:46:18Z", "t_script_end": "2026-09-24T02:49:18Z" }, "null_reasons": { "t_request": "not in closure.json: arm.py cannot observe the capacity request; the operator must supply it", "t_released": "not in closure.json: provider release is external to the seat (arm.py note: container stop is not release)" } }, "native_money": { "currency": "USD", "list_rate_per_gpu_hr": 2.99, "billing": "per-minute", "payer": "hotaisle-credit", "modeled_minutes": null, "modeled_minutes_basis": null, "modeled_usd": null, "modeled_lower_bound_minutes": 158.3, "modeled_lower_bound_usd": 7.8886, "modeled_lower_bound_basis": "t_ssh -> t_work_end at list; release and/or request unknown, so the billable window is at least this long", "billed_usd": null, "billed_ref": null, "credits_usd": null, "energy": { "watts_mean": null, "kwh": null, "tariff_usd_per_kwh": null, "usd": null, "meter": null, "reason": "cloud seat; no meter" }, "null_reasons": { "modeled_minutes": "t_request and/or t_released unknown (closure.json)", "modeled_minutes_basis": "see modeled_minutes", "modeled_usd": "buyer's window unknown; see modeled_lower_bound_usd", "billed_usd": "invoice not posted / not entered in closure.json", "billed_ref": "see billed_usd", "credits_usd": "not entered in closure.json (PREREG: credit, modeled and invoice are three fields)" } }, "evidence": { "ledger_sha256": "d0c57e6dcc731fdf9141ab8e66f6dcb30f612e31ae5a4a0dbf80eef1042f54fc", "detailed_sha256": "d20a5440675b122209a586b1e0353fe3ae957f0179183f66f85636370980a82a", "evaluation_sha256": "f715bc78a8fd9c35a2a9a904ba3a78dd8dea1da7cd103cbc45b7d52d95008893", "mapping_sha256": "e45613a58c11e21f795c887d1588dcc5e2989b02ffc02a0f693c0e7915b94994", "ledger_path": "campaign/results/run3-scored-a-t1/ledger/run3-run3-scored-a-t1.ledger.json", "directory": "campaign/results/run3-scored-a-t1" }, "evidence_scope": "Exact supplied files and judgment vector checked; outputs were not re-executed or independently graded.", "billing_status": "AWAITING_ALLOCATION_BILL", "whole_bill_share_usd": null }, { "run_id": "run3-scored-n-t0-20260924", "role": "scored_arm", "provider": "digitalocean", "allocation_id": "digitalocean-nyc2-603208981", "service_id": "gpu-h100x1-80gb", "region": "nyc2", "gpus": 1, "model": { "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8", "revision": "dcaee4d4dfc5ee71ad501f01f530e5652438fde0", "precision": "FP8", "bytes": 31187041238 }, "runtime": { "image": "vllm/vllm-openai", "digest": "sha256:8a69ffad015f138d7170c4ddc429e230a3bc1c1719f67e14324749df200a4b90", "engine": "vllm", "engine_version": "0.30.0", "attention_backend": [ "FLASH_ATTN" ], "linear_kernel": [ "FlashInferFp8DeepGEMMDynamicBlockScaledKernel" ] }, "workload_id": "sha256:6c508237082261ac927b96825e2f097782389b1cebc3f78360f0665a8db3501f", "counts": { "attempted": 8622, "completed": 8622, "correct": 4329, "accepted": 4280 }, "request_slots": 8622, "p95_ttft_including_queue_ms": 77.86753177642821, "p99_ttft_including_queue_ms": 2234.220626354125, "p95_e2e_including_queue_ms": 12068.458640575403, "acceptance_rule": { "basis": "run3/PREREG.md: accepted = completed AND EvalPlus base+plus pass AND first text <= 1 s AND end <= 60 s from the SCHEDULED arrival (queue_time + ttft, queue_time + latency in detailed.json)", "ttft_ms": 1000, "e2e_ms": 60000, "correctness": true, "queue": true }, "evaluator_criterion": "evalplus-0.3.1-human-v0.1.10-mbpp-v0.2.0-base-and-plus-raw-completion-v1", "clocks": { "t_request": "2026-09-24T04:08:03Z", "t_ssh": "2026-09-24T04:09:26Z", "t_ready": "2026-09-24T04:16:11Z", "t_work_start": "2026-09-24T04:16:16Z", "t_work_end": "2026-09-24T05:16:23Z", "t_released": "2026-09-24T05:17:52Z", "sources": { "t_request": { "source": "closure.json (operator: when the create/provision was requested)", "precision": "operator-recorded" }, "t_ssh": { "source": "arm.py --t-ssh (observed before the script started)", "precision": "1 s" }, "t_ready": { "source": "ledger-times.json: /v1/models listed the model", "precision": "1 s" }, "t_work_start": { "source": "ledger-times.json: replay child started", "precision": "1 s" }, "t_work_end": { "source": "ledger-times.json: replay child exited (or the finally-block stamp after an interrupt)", "precision": "1 s" }, "t_released": { "source": "closure.json (provider confirmation; container stop is not release)", "precision": "operator-recorded" }, "t_script_start": "2026-09-24T04:10:22Z", "t_script_end": "2026-09-24T05:16:24Z" }, "null_reasons": {} }, "native_money": { "currency": "USD", "list_rate_per_gpu_hr": 4.41, "billing": "per-second-5min-min", "payer": "operator", "modeled_minutes": 69.82, "modeled_minutes_basis": "t_request -> t_released from closure.json (the buyer's window)", "modeled_usd": 5.1315, "modeled_lower_bound_minutes": 69.82, "modeled_lower_bound_usd": 5.1315, "modeled_lower_bound_basis": "t_request -> t_released at list; equals the full window", "billed_usd": null, "billed_ref": null, "credits_usd": null, "energy": { "watts_mean": null, "kwh": null, "tariff_usd_per_kwh": null, "usd": null, "meter": null, "reason": "cloud seat; no meter" }, "null_reasons": { "billed_usd": "invoice not posted / not entered in closure.json", "billed_ref": "see billed_usd", "credits_usd": "not entered in closure.json (PREREG: credit, modeled and invoice are three fields)" } }, "evidence": { "ledger_sha256": "7903691bde1ee052d07e638f9db48264e28b74539e364d4bb4d350b14c9d9c20", "detailed_sha256": "f1a435b029675ffd2a7c6bb96fc0526c5294bb8acdda9ac76bbe764f9fb99b77", "evaluation_sha256": "7ff6854844b405ea067ec347750e80d8633087d65578c949f5d9144e58532349", "mapping_sha256": "599c9a8506c1d6065fac8c39ab461d67f235db7854ca8f7034eef75c226d8abf", "ledger_path": "campaign/results/run3-scored-n-t0/ledger/run3-run3-scored-n-t0.ledger.json", "directory": "campaign/results/run3-scored-n-t0" }, "evidence_scope": "Exact supplied files and judgment vector checked; outputs were not re-executed or independently graded.", "billing_status": "AWAITING_ALLOCATION_BILL", "whole_bill_share_usd": null }, { "run_id": "run3-smoke-a-t0-20260924", "role": "unscored_smoke", "provider": "hotaisle", "allocation_id": "hotaisle-enc1-gpuvm004", "service_id": "vm-mi300x-1x", "region": "enc1", "gpus": 1, "model": { "id": "Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8", "revision": "dcaee4d4dfc5ee71ad501f01f530e5652438fde0", "precision": "FP8", "bytes": 31187041238 }, "runtime": { "image": "vllm/vllm-openai-rocm:v0.30.0", "digest": "sha256:2e7da1ad1c66836802072588adea75f9f4991da5f9545b4318e91d422c22ce6a", "engine": "vllm", "engine_version": "0.30.0", "attention_backend": [ "ROCM_ATTN" ], "linear_kernel": [ "UNVERIFIED" ] }, "workload_id": "sha256:6c508237082261ac927b96825e2f097782389b1cebc3f78360f0665a8db3501f", "counts": { "attempted": 8622, "completed": 984, "correct": 491, "accepted": 458 }, "request_slots": 8622, "p95_ttft_including_queue_ms": 1629.8969984054563, "p99_ttft_including_queue_ms": 3375.768818855285, "p95_e2e_including_queue_ms": 11988.252842426293, "acceptance_rule": { "basis": "run3/PREREG.md: accepted = completed AND EvalPlus base+plus pass AND first text <= 1 s AND end <= 60 s from the SCHEDULED arrival (queue_time + ttft, queue_time + latency in detailed.json)", "ttft_ms": 1000, "e2e_ms": 60000, "correctness": true, "queue": true }, "evaluator_criterion": "evalplus-0.3.1-human-v0.1.10-mbpp-v0.2.0-base-and-plus-raw-completion-v1", "clocks": { "t_request": "2026-09-24T00:08:55Z", "t_ssh": "2026-09-24T00:10:57Z", "t_ready": "2026-09-24T00:18:34Z", "t_work_start": "2026-09-24T00:18:38Z", "t_work_end": "2026-09-24T00:27:51Z", "t_released": null, "sources": { "t_request": { "source": "closure.json (operator: when the create/provision was requested)", "precision": "operator-recorded" }, "t_ssh": { "source": "arm.py --t-ssh (observed before the script started)", "precision": "1 s" }, "t_ready": { "source": "ledger-times.json: /v1/models listed the model", "precision": "1 s" }, "t_work_start": { "source": "ledger-times.json: replay child started", "precision": "1 s" }, "t_work_end": { "source": "ledger-times.json: replay child exited (or the finally-block stamp after an interrupt)", "precision": "1 s" }, "t_released": { "source": "closure.json (provider confirmation; container stop is not release)", "precision": "operator-recorded" }, "t_script_start": "2026-09-24T00:12:03Z", "t_script_end": "2026-09-24T00:27:54Z" }, "null_reasons": { "t_released": "not in closure.json: provider release is external to the seat (arm.py note: container stop is not release)" } }, "native_money": { "currency": "USD", "list_rate_per_gpu_hr": 2.99, "billing": "per-minute", "payer": "hotaisle-credit", "modeled_minutes": null, "modeled_minutes_basis": null, "modeled_usd": null, "modeled_lower_bound_minutes": 18.93, "modeled_lower_bound_usd": 0.9435, "modeled_lower_bound_basis": "t_request -> t_work_end at list; release and/or request unknown, so the billable window is at least this long", "billed_usd": null, "billed_ref": null, "credits_usd": null, "energy": { "watts_mean": null, "kwh": null, "tariff_usd_per_kwh": null, "usd": null, "meter": null, "reason": "cloud seat; no meter" }, "null_reasons": { "modeled_minutes": "t_request and/or t_released unknown (closure.json)", "modeled_minutes_basis": "see modeled_minutes", "modeled_usd": "buyer's window unknown; see modeled_lower_bound_usd", "billed_usd": "invoice not posted / not entered in closure.json", "billed_ref": "see billed_usd", "credits_usd": "not entered in closure.json (PREREG: credit, modeled and invoice are three fields)" } }, "evidence": { "ledger_sha256": "3f397925db31ff6c53bb32dd857ed32cf5813bac783549a3dc5acfa9e6a726bd", "detailed_sha256": "bd1d1d490e783a69bcfdbe803595aed7ba016ccb0846c9a0132273b6bb3d8241", "evaluation_sha256": "c1fd70b70d7edc56a3b6a9def6f57e59b873ebfe639a3de2752cfaee72a7b279", "mapping_sha256": "94580451f7eab7c3776ccc3c3e63af6d8e168e88b2295332b51890ae20f59c1b", "ledger_path": "campaign/results/run3-smoke-a-t0/ledger/run3-run3-smoke-a-t0.ledger.json", "directory": "campaign/results/run3-smoke-a-t0" }, "evidence_scope": "Exact supplied files and judgment vector checked; outputs were not re-executed or independently graded.", "billing_status": "AWAITING_ALLOCATION_BILL", "whole_bill_share_usd": null } ], "allocation_bills": [], "unit_boundary": "One run remains one run. Requests and time buckets are not independent trial or provider counts.", "cost_boundary": "Realized whole-bill cost is separate from list-price modeled cost. v1.4 challenge total_cost_usd is not silently redefined.", "evidence_boundary": "Receipt hashes establish exact supplied bytes, not independent invoice authenticity or complete upstream history." }