Bielik v3 11B Instruct

GPAI Provider: SpeakLeash

Metadata

Model URL: https://huggingface.co/speakleash/Bielik-11B-v3.0-Instruct
When was this model published? 2025-12-31
Is this a new or a fine-tuned model? Fine-tuned model
When was the Public Summary last updated? 2026-01-01
When did we evaluate the Public Summary? 2026-01-12
Where did we find the Public Summary? We found this document linked on this page. An archived version can be found here.

Results of Evaluation

Bielik v3 11B Instruct	Transparency				Usefulness		General
	Clarity	Completeness	Consistency	Correctness	Accessibility	Comprehension	*Transparency*	*Usefulness*
Document	81.82	100.0	100.0	80.0	100.0	66.67	88.89	84.21
General information	78.89	86.59	100.0	100.0	100.0	90.91	87.03	93.75
Public Data Sources	100.0	74.29	100.0	100.0	100.0	100.0	86.91	100.0
Private Data Sources	N/A	100.0	100.0	100.0	N/A	N/A	100.0	N/A
Scraped/Crawled Data	91.6	81.48	100.0	77.88	0.0	67.95	83.78	51.46
User Data	N/A	100.0	100.0	100.0	N/A	N/A	100.0	N/A
Synthetic/Other Data	100.0	100.0	100.0	100.0	N/A	100.0	100.0	100.0
Data Processing	94.74	100.0	100.0	100.0	0.0	86.96	98.72	71.43
Sum	92.74	82.48	100.0	88.65	46.67	80.51	88.02	71.11
							B+	C+

Evaluation Notes

The open collaboration project SpeakLeash publishes a public summary for its Bielik v3 11B Instruct model on its website and also links it in its model card. The public summary follows the structure of the template, and describes Bielik as a fine-tuned version of Mistral. Of note, Bielik utilised the optional additional comments field in Section 1.3 to describe the purposes, regional, and linguistic characteristics of the dataset used as based on Polish and EU legal administrative domains and the inclusion of Silesian and Kashubian language Wikipedia pages. For their list of public data sources, however, the description referenced external documents ('preprints') which resulted in a deduction of points as relevant information was not provided in the document. We assessed its score to be 88.02% with grade B+ for transparency, and 71.11% with grade C+ for usefulness.