Qwen3-8B

MERA Created at 05.03.2026 00:11
0.645
The overall result
44
Place in the rating
Weak tasks:
120
RWSD
206
PARus
54
RCB
238
ruEthics
157
MultiQ
55
ruWorldTree
95
ruOpenBookQA
267
CheGeKa
99
ruMMLU
34
ruHateSpeech
117
ruDetox
149
ruHHH
66
ruTiE
26
ruHumanEval
111
USE
23
MathLogicQA
24
ruMultiAr
96
SimpleAr
32
LCS
120
BPS
60
ruModAr
110
MaMuRAMu
59
ruCodeEval
+19
Hide

Ratings for leaderboard tasks

The table will scroll to the left

Task name Result Metric
LCS 0.342 Accuracy
RCB 0.582 / 0.558 Accuracy F1 macro
USE 0.289 Grade norm
RWSD 0.588 Accuracy
PARus 0.844 Accuracy
ruTiE 0.856 Accuracy
MultiQ 0.456 / 0.296 F1 Exact match
CheGeKa 0.117 / 0.079 F1 Exact match
ruModAr 0.824 Exact match
MaMuRAMu 0.784 Accuracy
ruMultiAr 0.985 Exact match
ruCodeEval 0.371 / 0.458 / 0.482 Pass@k
MathLogicQA 0.99 Accuracy
ruWorldTree 0.983 / 0.788 Accuracy F1 macro
ruOpenBookQA 0.903 / 0.723 Accuracy F1 macro

Evaluation on open tasks:

Go to the ratings by subcategory

The table will scroll to the left

Task name Result Metric
BPS 0.973 Accuracy
ruMMLU 0.711 Accuracy
SimpleAr 0.99 Exact match
ruHumanEval 0.609 / 0.678 / 0.689 Pass@k
ruHHH 0.775
ruHateSpeech 0.86
ruDetox 0.273
ruEthics
Correct God Ethical
Virtue 0.293 0.361 0.401
Law 0.319 0.34 0.387
Moral 0.324 0.378 0.421
Justice 0.256 0.303 0.36
Utilitarianism 0.23 0.302 0.358

Information about the submission

Mera version
v1.2.0
Torch Version
2.9.1
The version of the codebase
7030e16
CUDA version
12.8
Precision of the model weights
bfloat16
Seed
1234
Batch
1
Transformers version
4.56.1
The number of GPUs and their type
1 x NVIDIA A100-SXM4-80GB
Architecture
local-chat-completions

Team:

MERA

Name of the ML model:

Qwen3-8B

Link to the ML model:

https://huggingface.co/Qwen/Qwen3-8B

Model size

8.0B

Model type:

Opened

SFT

Additional links:

https://arxiv.org/abs/2505.09388

Architecture description:

Qwen3-8B is a decoder-only transformer language model from the Qwen3 family with approximately 8 billion parameters. The model is designed for general language understanding and generation tasks, including reasoning, coding, and multilingual applications, and supports long-context inputs.

Description of the training:

The model is trained using a multi-stage pipeline consisting of large-scale pretraining followed by post-training. Post-training includes instruction tuning and alignment techniques to improve instruction following, reasoning ability, and response quality.

Pretrain data:

The model was pretrained on a large multilingual corpus of approximately 36 trillion tokens covering 119 languages. It was further post-trained on instruction-following and reasoning-oriented datasets to improve instruction following and reasoning performance.

License:

Apache License 2.0

Inference parameters

System prompt:
Реши задачу по инструкции ниже. Не давай никаких объяснений и пояснений к своему ответу. Не пиши ничего лишнего. Пиши только то, что указано в инструкции. Если по инструкции нужно решить пример, то напиши только числовой ответ без хода решения и пояснений. Если по инструкции нужно вывести букву, цифру или слово, выведи только его. Если по инструкции нужно выбрать один из вариантов ответа и вывести букву или цифру, которая ему соответствует, то выведи только эту букву или цифру, не давай никаких пояснений, не добавляй знаки препинания, только 1 символ в ответе. Если по инструкции нужно дописать код функции на языке Python, пиши сразу код, соблюдая отступы так, будто ты продолжаешь функцию из инструкции, не давай пояснений, не пиши комментарии, используй только аргументы из сигнатуры функции в инструкции, не пробуй считывать данные через функцию input. Не извиняйся, не строй диалог. Выдавай только ответ и ничего больше.

Description of the template:
{%- if tools %} {{- '<|im_start|>system \n' }} {%- if messages[0].role == 'system' %} {{- messages[0].content + ' \n \n' }} {%- endif %} {{- "# Tools \n \nYou may call one or more functions to assist with the user query. \n \nYou are provided with function signatures within <tools></tools> XML tags: \n<tools>" }} {%- for tool in tools %} {{- " \n" }} {{- tool | tojson }} {%- endfor %} {{- " \n</tools> \n \nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags: \n<tool_call> \n{"name": <function-name>, "arguments": <args-json-object>} \n</tool_call><|im_end|> \n" }} {%- else %} {%- if messages[0].role == 'system' %} {{- '<|im_start|>system \n' + messages[0].content + '<|im_end|> \n' }} {%- endif %} {%- endif %} {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %} {%- for message in messages[::-1] %} {%- set index = (messages|length - 1) - loop.index0 %} {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %} {%- set ns.multi_step_tool = false %} {%- set ns.last_query_index = index %} {%- endif %} {%- endfor %} {%- for message in messages %} {%- if message.content is string %} {%- set content = message.content %} {%- else %} {%- set content = '' %} {%- endif %} {%- if (message.role == "user") or (message.role == "system" and not loop.first) %} {{- '<|im_start|>' + message.role + ' \n' + content + '<|im_end|>' + ' \n' }} {%- elif message.role == "assistant" %} {%- set reasoning_content = '' %} {%- if message.reasoning_content is string %} {%- set reasoning_content = message.reasoning_content %} {%- else %} {%- if '</think>' in content %} {%- set reasoning_content = content.split('</think>')[0].rstrip(' \n').split('<think>')[-1].lstrip(' \n') %} {%- set content = content.split('</think>')[-1].lstrip(' \n') %} {%- endif %} {%- endif %} {%- if loop.index0 > ns.last_query_index %} {%- if loop.last or (not loop.last and reasoning_content) %} {{- '<|im_start|>' + message.role + ' \n<think> \n' + reasoning_content.strip(' \n') + ' \n</think> \n \n' + content.lstrip(' \n') }} {%- else %} {{- '<|im_start|>' + message.role + ' \n' + content }} {%- endif %} {%- else %} {{- '<|im_start|>' + message.role + ' \n' + content }} {%- endif %} {%- if message.tool_calls %} {%- for tool_call in message.tool_calls %} {%- if (loop.first and content) or (not loop.first) %} {{- ' \n' }} {%- endif %} {%- if tool_call.function %} {%- set tool_call = tool_call.function %} {%- endif %} {{- '<tool_call> \n{"name": "' }} {{- tool_call.name }} {{- '", "arguments": ' }} {%- if tool_call.arguments is string %} {{- tool_call.arguments }} {%- else %} {{- tool_call.arguments | tojson }} {%- endif %} {{- '} \n</tool_call>' }} {%- endfor %} {%- endif %} {{- '<|im_end|> \n' }} {%- elif message.role == "tool" %} {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %} {{- '<|im_start|>user' }} {%- endif %} {{- ' \n<tool_response> \n' }} {{- content }} {{- ' \n</tool_response>' }} {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} {{- '<|im_end|> \n' }} {%- endif %} {%- endif %} {%- endfor %} {%- if add_generation_prompt %} {{- '<|im_start|>assistant \n' }} {%- if enable_thinking is defined and enable_thinking is false %} {{- '<think> \n \n</think> \n \n' }} {%- endif %} {%- endif %}

Ratings by subcategory

Metric: Grade Norm
Model, team 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 8_0 8_1 8_2 8_3 8_4
Qwen3-8B
MERA
0.3 0.4 0.867 0.267 0.167 0.367 0.033 - 0 0 0.067 0 0.233 0.067 0.067 0.283 0.033 0.033 0 0.067 0.2 0.6 0.433 0.4 0.4 0.567 0.267 0.467 0.4 0.5 0.367
Model, team Honest Helpful Harmless
Qwen3-8B
MERA
0.738 0.78 0.81
Model, team Anatomy Virology Astronomy Marketing Nutrition Sociology Management Philosophy Prehistory Human aging Econometrics Formal logic Global facts Jurisprudence Miscellaneous Moral disputes Business ethics Biology (college) Physics (college) Human Sexuality Moral scenarios World religions Abstract algebra Medicine (college) Machine learning Medical genetics Professional law PR Security studies Chemistry (школьная) Computer security International law Logical fallacies Politics Clinical knowledge Conceptual_physics Math (college) Biology (high school) Physics (high school) Chemistry (high school) Geography (high school) Professional medicine Electrical engineering Elementary mathematics Psychology (high school) Statistics (high school) History (high school) Math (high school) Professional accounting Professional psychology Computer science (college) World history (high school) Macroeconomics Microeconomics Computer science (high school) European history Government and politics
Qwen3-8B
MERA
0.615 0.524 0.855 0.808 0.739 0.741 0.825 0.653 0.664 0.691 0.667 0.778 0.58 0.704 0.775 0.63 0.73 0.847 0.844 0.84 0.459 0.754 0.89 0.751 0.679 0.82 0.413 0.62 0.686 0.63 0.76 0.777 0.675 0.808 0.743 0.868 0.78 0.894 0.841 0.867 0.798 0.75 0.766 0.958 0.884 0.829 0.784 0.944 0.596 0.675 0.82 0.793 0.838 0.895 0.87 0.782 0.819
Model, team SIM FL STA
Qwen3-8B
MERA
0.779 0.52 0.707
Model, team Anatomy Virology Astronomy Marketing Nutrition Sociology Managment Philosophy Pre-History Gerontology Econometrics Formal logic Global facts Jurisprudence Miscellaneous Moral disputes Business ethics Bilology (college) Physics (college) Human sexuality Moral scenarios World religions Abstract algebra Medicine (college) Machine Learning Genetics Professional law PR Security Chemistry (college) Computer security International law Logical fallacies Politics Clinical knowledge Conceptual physics Math (college) Biology (high school) Physics (high school) Chemistry (high school) Geography (high school) Professional medicine Electrical Engineering Elementary mathematics Psychology (high school) Statistics (high school) History (high school) Math (high school) Professional Accounting Professional psychology Computer science (college) World history (high school) Macroeconomics Microeconomics Computer science (high school) Europe History Government and politics
Qwen3-8B
MERA
0.6 0.851 0.8 0.667 0.855 0.672 0.707 0.667 0.654 0.754 0.808 0.8 0.567 0.744 0.743 0.654 0.748 0.867 0.842 0.807 0.649 0.847 0.844 0.852 0.867 0.848 0.795 0.719 0.895 0.889 0.867 0.795 0.75 0.825 0.636 0.821 0.933 0.867 0.895 0.815 0.845 0.762 0.844 0.978 0.914 0.889 0.879 0.909 0.785 0.895 0.844 0.739 0.823 0.766 0.721 0.637 0.844
Coorect
Good
Ethical
Model, team Virtue Law Moral Justice Utilitarianism
Qwen3-8B
MERA
0.293 0.319 0.324 0.256 0.23
Model, team Virtue Law Moral Justice Utilitarianism
Qwen3-8B
MERA
0.361 0.34 0.378 0.303 0.302
Model, team Virtue Law Moral Justice Utilitarianism
Qwen3-8B
MERA
0.401 0.387 0.421 0.36 0.358
Model, team Women Men LGBT Nationalities Migrants Other
Qwen3-8B
MERA
0.87 0.8 0.941 0.838 0.857 0.869