Arcbright Elf 1 12B (r2)v1.7 · 10 s rule
Data: JevBench v1.7.14, published 2026-10-28 · method v1.7 (10 s rule), measured under the rule
Capability 79.9 · Not ranked in the current release
Measured on the 27 Oct 2026 draw, which is not yet equated to the reference scale. This draw differs strongly from earlier draws (72% multilingual items). Shown unranked until the draw is equated.
Last measured: 2026-10-27 · measurement release: v1.7.14
Apache-2.0 (model card, root LICENSE and serve/LICENSE; base google/gemma-4-12B-it under the Gemma 4 licence), public, not gated
intelligence
66.8
calibration
92.9
speed
92.5
cost
60.6
Composite (secondary)
75.4 · Not ranked on this board
Cost and measurement conditions
$0.021 per 1,000 decisions (estimated)
ESTIMATE (*est.), self-hosted open weights, no public tariff: the declared base google/gemma-4-12B-it has no listing in the frozen pricing module (sha256 645fd82eac728d597048bec0995ecd3ac896566e269fb6c9a1fdb84c171a9ac4), so the settled Gemma-4-12B estimate applies: the nearest listed 12B Gemma sibling, google/gemma-3-12b-it at USD 0.05/0.15 per million input/output tokens, the same explicit reference as the earlier google/gemma-4-12B-it rows. Tokens as reported by the author's server usage block (the full rendered chat prompt incl. system prompt, options and answer prefill; equal to our offline count on every item): 610,840 input tokens (the 5 cut items count at their cut size, 30,789 tokens) and 1,500 output tokens (one forward pass per answered question, reported as 1 output token each). COST FIGURES: USD 0.020511 per 1,000 decisions over all 1,500 items of the draw (every row on this draw is costed on all 1,500 items); 0.32x the official cost cap USD 0.064595. Estimated model tariff, not charged usage or a host quote.
Measured p50 latency: 0.029 s (on-GPU, no network, loopback; not directly comparable with API latency, which we time end to end from our measuring host, network included)
Ranking and cap use: 0.21 s (est. incl. network (2x + 150 ms))
Self-hosted model: measured p50 0.029 s on-GPU, no network (timed on our GPU over loopback). Not directly comparable with API latency, which we time end to end from our measuring host, network included. The ranking and latency cap use the same estimate for every self-hosted row: 2 × measured + 150 ms = 0.21 s. Model authors can ask us to time a hosted endpoint instead (on request).
x2 + 0.15 s (assumption, not measured)
Evaluator-owned rented Verda GPU VM, one NVIDIA RTX PRO 6000 Blackwell Server Edition 96 GB (97,887 MiB, compute capability 12.0, driver 580.178.04), FIN-03 (Finland, EU), PAYG non-spot, instance type 1RTXPRO6000.30V, image 24.04.cuda13.0, USD 2.608 per hour recorded by the runner, 150 GB disk. One VM for this row, one attempt. Measured 27 Oct 2026, VM window about 17:54-18:06Z, about USD 0.5 GPU rental (elapsed time at the recorded hourly rate). Author server at 127.0.0.1:8000 inside a network-none Docker container (cap-drop ALL), network cut proven from inside (loopback only, no default route, DNS and TCP fail), offline and credential-free, one request at a time, private output. Before sealed input: model files verified, the Triton gate, the smoke and a pre-warm that paid Triton's per-shape compilation passed, and the TTL fit gate admitted the run. Input sha256 9b2f32edaccb29afd726af8cd54cf1a22d85e8b1824566924045d83c6f828426 (v1.6-regular-20261027) was re-checked inside the container (match). Complete 1,500-row run, rc 0. Raw output sha256 84f08ebb6ff92af453b75cda042f617ea89c5be6a2176c7c69e11f2521da6f08 (the pull-hash receipt). Node-shred and local-shred receipts verified; exact-id teardown receipt confirms the VM closed. GPU rental is separate from the benchmark's estimated model tariff.
Measured latency over 1,500 decisions: p50 0.029 s · p95 0.060 s
0 decisions over 10 s.
A decision that takes longer than 10 seconds (measured time) counts as a wrong answer and as zero calibration for that item: each type's calibration is multiplied by the share of its decisions answered within 10 s. The limit uses the measured time (raw). The displayed adjusted time for self-hosted rows (2 × + 0.15 s) is an assumption and is not used for the limit.