Messier One v0.4 (Qwen3.5-4B fine-tune, self-hosted)v1.7 · 10 s rule
Data: JevBench v1.8.15, published 2026-11-17 · method v1.7 (10 s rule), re-scored from the recorded per-decision times
Capability 62.4 · rank #120 of 243 on the open-weights board
Last measured: 2026-10-14 · measurement release: v1.6.22
Weights CC BY-NC 4.0 (non-commercial: free for research and evaluation, commercial licence on request; listing a measured score is evaluation use); code Apache-2.0; base Qwen/Qwen3.5-4B Apache-2.0
intelligence
52.0
calibration
72.9
speed
89.4
cost
75.7
Composite (secondary)
69.8 · Composite rank #79
Cost and measurement conditions
$0.0065 per 1,000 decisions (estimated)
GPU-time estimate from v1.8.0: one monthly-rented RTX PRO 6000 96 GB, EUR 999/month / 730 hours, ECB USD 1.1133/EUR dated 28 Oct 2026, USD 1.5235/hour; average load 70%. Architecture: small qwen, 4B active parameters; throughput 61525 tokens/s (architecture-class calibration, 29 Oct 2026, efficient-engine prefill; capped at 100,000 tokens/s). The row's own scored cost basis: 978,475 input + 0 output tokens over 1,485 decisions = 658.905723906 tokens/decision. GPU-time cost USD 0.00647460943332089/1,000; old list-price cost USD 0.0197671717171717/1,000; applied min(old list-price cost, GPU-time cost) = USD 0.00647460943332089/1,000. The GPU-time estimate is lower. Estimated, not a bill. Intelligence, Calibration, measured latency and the common cost-item membership are unchanged. Prior prices remain in the archived v1.7 releases.
Measured p50 latency: 0.069 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: 0.29 s (est. incl. network (2x + 150 ms))
Self-hosted model: measured p50 0.069 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 = 0.29 s. Model authors can ask us to time a hosted endpoint instead (on request).
x2 + 0.15 s (assumption, not measured)
Evaluator-owned rented Lium GPU pod (measurement receipts on file), 1x NVIDIA RTX A6000 48 GB - different hardware from the H100/RTX 5090 rows of this draw and from the RTX 3090 the author quotes. Started by the author's OWN ./serve_vllm.sh MODEL_DIR 8080 with its defaults, which is what the author insists on for v0.4: vLLM 0.30.0, --max-model-len 8192, chunked prefill and prefix caching OFF (this revision is non-causal), --max-logprobs 256, processed_logprobs, front end --max-state 7500 --max-total 8150 --gate 0.5 --noul-floor 0.801; transformers 5.18.0 exactly as pinned. One request at a time from the unchanged runner. Isolation: the server ran inside a network namespace with loopback only (unshare -n), the cut proven before the sealed input was staged and re-proven immediately before the first sealed request, with the synthetic smoke gate passed in between; a network namespace replaces docker --network none because the pod carried no local image (under the pre-registration). The pod host kept its network for ssh.
Measured latency over 1,500 decisions: p50 0.069 s · p95 0.123 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.
JevBench v1.6.43 values (method v1.6, before the 10 s rule)
Capability 62.44 · Composite 66.16 · rank #45 on the all-systems board · #24 open weights
intelligence 52.03 · calibration 72.86 · speed 89.67 · cost 61.12
v1.6 speed basis (124 items): p50 0.066 s · p95 0.116 s
These figures are kept for comparison. The current figures above are re-scored under the 10 s rule from the recorded per-decision times and replace them.