Loading comparison…
HEAD TO HEAD
Yuichi Tateno (hotchpotch): bekko-system-one and Nandakishor M: vega, compared on what they decide, where they run, what they cost and what their publishers report.
| Property | hotchpotch/bekko-system-one | nandakishorm/vega |
|---|---|---|
| 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. | Nandakishor M's typed-decision model. A frozen Qwen3.5 backbone is read, never trained, and cut off mid-stack; a small trained physics engine settles each choice, score or yes/no question into calibrated probabilities, a conformal set and an abstain flag. 0.8B and 4B. |
| Decides | choice, noul, score | choice, score, noul, classify, route |
| Architecture | bekko | vega |
| Fine-tuned from | cross-encoder/ettin-reranker-400m-v1 | qwen/qwen3.5-0.8b |
| 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 | — | 55.6% |
| Calibration error | — | 0.151 |
| 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) | His 11 October 2026 benchmark, published with its results and its harness (a notebook in vegaml): macro accuracy over 43 public tasks for the 0.8B, 0.717 for the 4B; with JevBench (231 items) and DecisionBench (1,075 rows), 3,841 decisions in all, seed 20261011. On his own set of 2,050 choice decisions over six intent workflows (esci, clinc150, appliances, banking77, when2call, apibank), one choice question each, which he has not published with its question wording, he reports 0.763 (calibration error 0.026) for the 0.8B with its public_intents adapter, 0.389 without the adapter and 0.803 for the 4B with it. He publishes no accuracy for the typed_decisions adapters |
| Latest version | 2026.09 | 2026.10.12 |
| Variants | 0_BekkoInference | adapters, backbone, 4b |
| Size of latest version | 1.5 GB | 10.5 GB |
| Files | 9 | 31 |
| Downloads | 0 | 108 |
| Stars | 0 | 0 |
| Tags | system-one, encoder, ettin, onnx, browser, small, 395m | system-one, qwen, frozen-backbone, calibrated, conformal, abstention, adapters, multimodal, long-context, 0.8b |
| Updated | Oct 7, 2026 | Oct 29, 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 vega from Nandakishor M. Both have open weights you can download and run. Both answer choice, noul and score questions. Only vega answers classify and route. bekko-system-one is the smaller model, at 395M parameters to 800M. bekko-system-one is licensed other; vega, apache-2.0.
Only vega publishes an accuracy figure (55.6% on His 11 October 2026 benchmark, published with its results and its harness (a notebook in vegaml): macro accuracy over 43 public tasks for the 0.8B, 0.717 for the 4B; with JevBench (231 items) and DecisionBench (1,075 rows), 3,841 decisions in all, seed 20261011. On his own set of 2,050 choice decisions over six intent workflows (esci, clinc150, appliances, banking77, when2call, apibank), one choice question each, which he has not published with its question wording, he reports 0.763 (calibration error 0.026) for the 0.8B with its public_intents adapter, 0.389 without the adapter and 0.803 for the 4B with it. He publishes no accuracy for the typed_decisions adapters); bekko-system-one does not, so there is no comparison to make without your own test.
bekko-system-one: Free (open weights). vega: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Yes, both: systemone pull hotchpotch/bekko-system-one and systemone pull nandakishorm/vega download the weights.