s1-code v1

First version of s1-code. For use, prefer v3. A System One decision model (322M, CPU) that returns the probability that a Python function answers a search. Fine-tuned from convaiinnovations/laya (Apache 2.0).

Same input format as v3: the state is the file path, the function name, an empty line and the first 1,500 characters of the source; the question is the noul This code answers the search: <your search>.

Results (new held-out test, 197 questions, 25 candidates from Qwen3-Embedding)

System Top 1 English Spanish
s1-code v1 alone 144 68 % 78 %
s1-code v1 + Qwen3-Embedding (fused) 167 79 % 91 %
Qwen3-Embedding alone 152 76 % 78 %

Trained on 30,068 generated questions (59 % Spanish, 41 % English) for functions of 306 public Python repositories with permissive licences, with granite-mined negatives. Stronger in Spanish than in English, which v2 and v3 corrected.

Licence

Apache 2.0. Fine-tuned from Laya by Convai Innovations.

Downloads last month

-

Downloads are not tracked for this model. How to track
Safetensors
Model size
0.3B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for api-service-sac/s1-code-v1

Finetuned
(134)
this model