Bosun v3.1 0.6B
system-one-open · by Clause Logic · Code and weights marked open
Base model: Qwen/Qwen3-0.6Bsourcev1.5 roster addendum A4
apache-2.0, not gated; LICENSE + NOTICE (Clause Logic Inc.) present at the pinned revision
JevBench v1.5.4 score
2.500
Option A: rank #73 of 106 ranked systems.
The three option scores and ranks are published independently; the headline is Option A.
| Option | Score | Rank |
|---|---|---|
| A · headline | 2.500 | #73 of 106 |
| B | 1.924 | #73 of 106 |
| C | 1.736 | #73 of 106 |
Published axes
- intelligence
- 13.6
- calibration
- 65.2
- speed
- 61.8
- cost
- 77.5
Bands use the published 0–100 axis values; the marked reference is Jev 1.13.0 when that axis is available.
Run and cost evidence
- Run status
- complete · 1,624 decisions · 0 missing
- Cost
- $0.0056 per 1,000 decisions · estimate · ESTIMATE: documented hosted-model estimate; no exact base-model floor applies. $0.01/M input, $0/M output; same-base frozen raw-qwen3-0.6b estimate.
- Median latency
- 4.037 seconds, adjusted · x2 + 0.15 s (assumption, not measured)
- Endpoint condition
- evaluator-owned CPU, Sandy (AMD Ryzen 5 3600), 4 of 12 threads, nice -n 5, shared host, offline (HF_HUB_OFFLINE=1, local_files_only=True), credential-free (no API key, no HF token)
- Published source
- https://huggingface.co/Hanno-Labs/bosun-v3.1-0.6b
Model and serving disclosure
Hanno-Labs/bosun-v3.1-0.6b @ 1d8b6f9611f9b64b514ce8b57cd86398fbc31a3b; base Qwen/Qwen3-0.6B @ c1899de289a04d12100db370d81485cdf75e47ca. Author predict interface preserves candidate mapping and masking. Offline shared AMD Ryzen 5 3600 CPU, four threads; float32 instead of dtype=auto/bfloat16. This numerical precision deviation does not establish parity with bfloat16. Shared-host latency is not a controlled GPU comparison. The 1.7B sibling is not included. The public/sealed gap and frozen overfit penalty remain applied.
Values come from the public v1.5.4 aggregate. Scores and ranks may change in a later release.
Read the full leaderboard, the v1.5.4 release page, and the published method.