license, language, base_model, datasets, model-index

license	language	base_model	datasets	model-index
cc-by-nc-4.0	ro	google/gemma-7b	OpenLLM-Ro/ro_sft_alpaca OpenLLM-Ro/ro_sft_alpaca_gpt4 OpenLLM-Ro/ro_sft_dolly OpenLLM-Ro/ro_sft_selfinstruct_gpt4 OpenLLM-Ro/ro_sft_norobots OpenLLM-Ro/ro_sft_orca OpenLLM-Ro/ro_sft_camel OpenLLM-Ro/ro_sft_oasst OpenLLM-Ro/ro_sft_ultrachat OpenLLM-Ro/ro_sft_magpie_mt OpenLLM-Ro/ro_sft_magpie_reasoning	name results OpenLLM-Ro/RoGemma-7b-Instruct-2025-04-23 task dataset metrics type text-generation name type RoMT-Bench RoMT-Bench name type value Score Score 6.28 task dataset metrics type text-generation name type RoCulturaBench RoCulturaBench name type value Score Score 3.65 task dataset metrics type text-generation name type Romanian_Academic_Benchmarks Romanian_Academic_Benchmarks name type value Average accuracy accuracy 50.52 task dataset metrics type text-generation name type OpenLLM-Ro/ro_arc_challenge OpenLLM-Ro/ro_arc_challenge name type value Average accuracy accuracy 47.70 task dataset metrics type text-generation name type OpenLLM-Ro/ro_mmlu OpenLLM-Ro/ro_mmlu name type value Average accuracy accuracy 51.66 task dataset metrics type text-generation name type OpenLLM-Ro/ro_winogrande OpenLLM-Ro/ro_winogrande name type value Average accuracy accuracy 66.32 task dataset metrics type text-generation name type OpenLLM-Ro/ro_hellaswag OpenLLM-Ro/ro_hellaswag name type value Average accuracy accuracy 53.59 task dataset metrics type text-generation name type OpenLLM-Ro/ro_gsm8k OpenLLM-Ro/ro_gsm8k name type value Average accuracy accuracy 36.04 task dataset metrics type text-generation name type OpenLLM-Ro/ro_truthfulqa OpenLLM-Ro/ro_truthfulqa name type value Average accuracy accuracy 47.81 task dataset metrics type text-generation name type LaRoSeDa_binary LaRoSeDa_binary name type value Average macro-f1 macro-f1 95.44 task dataset metrics type text-generation name type LaRoSeDa_multiclass LaRoSeDa_multiclass name type value Average macro-f1 macro-f1 59.24 task dataset metrics type text-generation name type WMT_EN-RO WMT_EN-RO name type value Average bleu bleu 25.17 task dataset metrics type text-generation name type WMT_RO-EN WMT_RO-EN name type value Average bleu bleu 21.17 task dataset metrics type text-generation name type XQuAD XQuAD name type value Average exact_match exact_match 15.88 task dataset metrics type text-generation name type XQuAD XQuAD name type value Average f1 f1 29.16 task dataset metrics type text-generation name type STS STS name type value Average spearman spearman 75.90 task dataset metrics type text-generation name type STS STS name type value Average pearson pearson 75.16 task dataset metrics type text-generation name type RoMT-Bench RoMT-Bench name type value First turn Score 6.97 name type value Second turn Score 5.58 task dataset metrics type text-generation name type OpenLLM-Ro/ro_arc_challenge OpenLLM-Ro/ro_arc_challenge name type value 0-shot accuracy 46.19 name type value 1-shot accuracy 46.53 name type value 3-shot accuracy 46.02 name type value 5-shot accuracy 48.33 name type value 10-shot accuracy 49.27 name type value 25-shot accuracy 49.87 task dataset metrics type text-generation name type OpenLLM-Ro/ro_mmlu OpenLLM-Ro/ro_mmlu name type value 0-shot accuracy 51.13 name type value 1-shot accuracy 50.94 name type value 3-shot accuracy 52.67 name type value 5-shot accuracy 51.90 task dataset metrics type text-generation name type OpenLLM-Ro/ro_winogrande OpenLLM-Ro/ro_winogrande name type value 0-shot accuracy 67.40 name type value 1-shot accuracy 65.04 name type value 3-shot accuracy 65.67 name type value 5-shot accuracy 67.17 task dataset metrics type text-generation name type OpenLLM-Ro/ro_hellaswag OpenLLM-Ro/ro_hellaswag name type value 0-shot accuracy 58.03 name type value 1-shot accuracy 56.63 name type value 3-shot accuracy 52.47 name type value 5-shot accuracy 48.63 name type value 10-shot accuracy 52.18 task dataset metrics type text-generation name type OpenLLM-Ro/ro_gsm8k OpenLLM-Ro/ro_gsm8k name type value 1-shot accuracy 24.11 name type value 3-shot accuracy 37.76 name type value 5-shot accuracy 46.25 task dataset metrics type text-generation name type LaRoSeDa_binary LaRoSeDa_binary name type value 0-shot macro-f1 96.33 name type value 1-shot macro-f1 94.62 name type value 3-shot macro-f1 95.06 name type value 5-shot macro-f1 95.76 task dataset metrics type text-generation name type LaRoSeDa_multiclass LaRoSeDa_multiclass name type value 0-shot macro-f1 43.65 name type value 1-shot macro-f1 64.30 name type value 3-shot macro-f1 64.22 name type value 5-shot macro-f1 64.81 task dataset metrics type text-generation name type WMT_EN-RO WMT_EN-RO name type value 0-shot bleu 13.30 name type value 1-shot bleu 28.59 name type value 3-shot bleu 29.48 name type value 5-shot bleu 29.31 task dataset metrics type text-generation name type WMT_RO-EN WMT_RO-EN name type value 0-shot bleu 1.11 name type value 1-shot bleu 18.97 name type value 3-shot bleu 31.99 name type value 5-shot bleu 32.60 task dataset metrics type text-generation name type XQuAD_EM XQuAD_EM name type value 0-shot exact_match 17.31 name type value 1-shot exact_match 12.44 name type value 3-shot exact_match 13.11 name type value 5-shot exact_match 20.67 task dataset metrics type text-generation name type XQuAD_F1 XQuAD_F1 name type value 0-shot f1 29.90 name type value 1-shot f1 24.24 name type value 3-shot f1 25.64 name type value 5-shot f1 36.86 task dataset metrics type text-generation name type STS_Spearman STS_Spearman name type value 1-shot spearman 76.50 name type value 3-shot spearman 73.63 name type value 5-shot spearman 77.58 task dataset metrics type text-generation name type STS_Pearson STS_Pearson name type value 1-shot pearson 75.15 name type value 3-shot pearson 72.69 name type value 5-shot pearson 77.63

Model Card for Model ID

This model points/is identical to RoGemma-7b-Instruct-2025-04-23.

RoGemma is a family of pretrained and fine-tuned generative text models for Romanian. This is the repository for the instruct 7B model. Links to other models can be found at the bottom of this page.

Model Details

Model Description

OpenLLM-Ro represents the first open-source effort to build a LLM specialized for Romanian. OpenLLM-Ro developed and publicly releases a collection of Romanian LLMs, both in the form of foundational model and instruct and chat variants.

• Developed by: OpenLLM-Ro

• Language(s): Romanian
• License: cc-by-nc-4.0
• Finetuned from model: gemma-7b
• Trained using: RoAlpaca, RoAlpacaGPT4, RoDolly, RoSelfInstruct, RoNoRobots, RoOrca, RoCamel, RoOpenAssistant, RoUltraChat, RoMagpiePro, RoMagpieReasoning

Model Sources

• Repository: https://github.com/OpenLLM-Ro/LLaMA-Factory
• Paper: https://arxiv.org/abs/2406.18266

Intended Use

Intended Use Cases

RoGemma is intented for research use in Romanian. Base models can be adapted for a variety of natural language tasks while instruction and chat tuned models are intended for assistant-like chat.

Out-of-Scope Use

Use in any manner that violates the license, any applicable laws or regluations, use in languages other than Romanian.

How to Get Started with the Model

Use the code below to get started with the model.

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OpenLLM-Ro/RoGemma-7b-Instruct")
model = AutoModelForCausalLM.from_pretrained("OpenLLM-Ro/RoGemma-7b-Instruct")

instruction = "Ce jocuri de societate pot juca cu prietenii mei?"
chat = [
        {"role": "user", "content": instruction},
        ]
prompt = tokenizer.apply_chat_template(chat, tokenize=False, system_message="")

inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt")
outputs = model.generate(input_ids=inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0]))

Academic Benchmarks

Model	Average	ARC	MMLU	Winogrande	Hellaswag	GSM8k	TruthfulQA
gemma-1.1-7b-it	41.44	40.32	47.22	55.01	47.03	9.50	49.58
RoGemma-7b-Instruct-2024-06-28	53.41	52.44	54.44	69.36	61.96	31.06	51.23
RoGemma-7b-Instruct-2024-10-09	50.48	52.01	52.37	66.97	56.34	25.98	49.18
RoGemma-7b-Instruct-2025-04-23	50.52	47.70	51.66	66.32	53.59	36.04	47.81
RoGemma-7b-Instruct-DPO-2024-10-09	48.27	46.66	54.45	63.73	49.33	34.98	40.45

Downstream tasks

	LaRoSeDa				WMT
	Few-shot		Finetuned		Few-shot		Finetuned
Model	Binary (Macro F1)	Multiclass (Macro F1)	Binary (Macro F1)	Multiclass (Macro F1)	EN-RO (Bleu)	RO-EN (Bleu)	EN-RO (Bleu)	RO-EN (Bleu)
gemma-1.1-7b-it	87.54	51.48	83.87	85.61	17.96	27.74	25.48	36.11
RoGemma-7b-Instruct-2024-06-28	97.86	65.70	98.43	87.17	27.91	23.08	27.99	39.51
RoGemma-7b-Instruct-2024-10-09	86.96	56.72	98.80	85.81	24.45	14.20	25.96	39.07
RoGemma-7b-Instruct-2025-04-23	95.44	59.24	-	-	25.17	21.17	-	-
RoGemma-7b-Instruct-DPO-2024-10-09	96.45	63.23	-	-	20.73	7.87	-	-

	XQuAD				STS
	Few-shot		Finetuned		Few-shot		Finetuned
Model	(EM)	(F1)	(EM)	(F1)	(Spearman)	(Pearson)	(Spearman)	(Pearson)
gemma-1.1-7b-it	42.10	62.30	60.34	77.40	49.10	50.23	83.43	83.64
RoGemma-7b-Instruct-2024-06-28	17.75	28.11	52.02	68.43	73.96	75.16	86.45	86.31
RoGemma-7b-Instruct-2024-10-09	26.03	41.58	46.72	60.79	73.23	71.58	88.42	88.45
RoGemma-7b-Instruct-2025-04-23	15.88	29.16	-	-	75.90	75.16	-	-
RoGemma-7b-Instruct-DPO-2024-10-09	19.14	38.10	-	-	69.38	69.34	-	-

MT-Bench

Model	Average	1st turn	2nd turn	Answers in Ro
gemma-1.1-7b-it	4.83	5.11	4.55	160/160
RoGemma-7b-Instruct-2024-06-28	5.26	5.92	4.60	160/160
RoGemma-7b-Instruct-2024-10-09	5.24	5.55	4.94	160/160
RoGemma-7b-Instruct-2025-04-23	6.28	6.97	5.58	160/160
RoGemma-7b-Instruct-DPO-2024-10-09	5.47	5.92	5.03	160/160

RoCulturaBench

Model	Average	Answers in Ro
gemma-1.1-7b-it	3.38	100/100
RoGemma-7b-Instruct-2024-06-28	3.26	100/100
RoGemma-7b-Instruct-2024-10-09	3.51	100/100
RoGemma-7b-Instruct-2025-04-23	3.65	100/100
RoGemma-7b-Instruct-DPO-2024-10-09	3.94	100/100

RoGemma Model Family

Model	Link
RoGemma-7b-Instruct-2024-06-28	link
RoGemma-7b-Instruct-2024-10-09	link
RoGemma-7b-Instruct-2025-04-23	link
RoGemma-7b-Instruct-DPO-2024-10-09	link

Citation

@inproceedings{masala-etal-2024-vorbesti,
    title = "``Vorbe\c{s}ti Rom{\^a}ne\c{s}te?'' A Recipe to Train Powerful {R}omanian {LLM}s with {E}nglish Instructions",
    author = "Masala, Mihai and Ilie-Ablachim, Denis and Dima, Alexandru and Corlatescu, Dragos Georgian and Zavelca, Miruna-Andreea and Olaru, Ovio and Terian, Simina-Maria and Terian, Andrei and Leordeanu, Marius and Velicu, Horia and Popescu, Marius and Dascalu, Mihai and Rebedea, Traian",
    editor = "Al-Onaizan, Yaser and Bansal, Mohit and Chen, Yun-Nung",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024",
    month = nov,
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.findings-emnlp.681/",
    doi = "10.18653/v1/2024.findings-emnlp.681",
    pages = "11632--11647"
}