HEAD TO HEAD
Yuichi Tateno (hotchpotch): bekko-system-one and TokenRhythm: neohorse-jev, compared on what they decide, where they run, what they cost and what their publishers report.
| Property | hotchpotch/bekko-system-one | tokenrhythm/neohorse-jev |
|---|---|---|
| Summary | Ultra-small English encoders (17M, 68M and 395M) fine-tuned from Ettin rerankers. Each answers choice, noul and score questions with a probability per candidate and no generated text. Experimental v0 by one developer; the weights have no licence assigned yet. | TokenRhythm's prefill-only decision model for agent workflows. Built on its NeoHorse-1-4B (a Qwen3.5-4B derivative), it predicts Choice, Noul and Score answers over application-defined options without generating text, with optional single-image input. |
| Decides | choice, noul, score | choice, score, noul, classify, route |
| Architecture | bekko | neohorse |
| Fine-tuned from | cross-encoder/ettin-reranker-400m-v1 | tokenrhythm/neohorse-1-4b |
| License | Unassigned: the weights licence is not finalised (treated as non-commercial); training code MIT | apache-2.0 |
| Availability | Open weights | Open weights |
| Hosted by | — | — |
| Input price | — | — |
| Decision accuracy | — | 75.3% |
| Calibration error | — | — |
| Valid action rate | — | — |
| Median latency | 5 ms | — |
| p95 latency | 5.4 ms | — |
| Evaluation suite | The maker's short-input speed probe (4-option Choice, RTX 5090, 1 Oct 2026) | JevBench public set (231 items), vLLM, maker's run |
| Latest version | 2026.09 | 2026.09 |
| Variants | 0_BekkoInference | backbone, SHA256SUMS, assets, dist, package, tokenizer, vision |
| Size of latest version | 1.5 GB | 8.5 GB |
| Files | 9 | 38 |
| Downloads | 0 | 3 |
| Stars | 0 | 0 |
| Tags | system-one, encoder, ettin, onnx, browser, small, 395m | system-one, qwen, multimodal, vllm, sglang, 4b |
| Updated | Oct 7, 2026 | Oct 7, 2026 |
Figures are from each model’s manifest; accuracy and latency are what the publishers report, on their own suites and hardware. Add a third model.
bekko-system-one is from Yuichi Tateno (hotchpotch) and neohorse-jev from TokenRhythm. Both have open weights you can download and run. Both answer choice, noul and score questions. Only neohorse-jev answers classify and route. bekko-system-one is the smaller model, at 395M parameters to 4.0B. bekko-system-one is licensed other; neohorse-jev, apache-2.0.
Only neohorse-jev publishes an accuracy figure (75.3% on JevBench public set (231 items), vLLM, maker's run); bekko-system-one does not, so there is no comparison to make without your own test.
bekko-system-one: Free (open weights). neohorse-jev: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Yes, both: systemone pull hotchpotch/bekko-system-one and systemone pull tokenrhythm/neohorse-jev download the weights.