Qwen3-0.6B

MERA Created at 04.03.2026 14:36
0.394
The overall result
340
Place in the rating
Weak tasks:
529
RWSD
445
PARus
413
RCB
478
ruEthics
579
MultiQ
420
ruWorldTree
452
ruOpenBookQA
513
CheGeKa
444
ruMMLU
521
ruHateSpeech
503
ruDetox
479
ruHHH
379
ruTiE
198
ruHumanEval
375
USE
41
MathLogicQA
36
ruMultiAr
418
SimpleAr
363
LCS
473
BPS
387
ruModAr
446
MaMuRAMu
154
ruCodeEval
+19
Hide

Ratings for leaderboard tasks

The table will scroll to the left

Task name Result Metric
LCS 0.092 Accuracy
RCB 0.473 / 0.399 Accuracy F1 macro
USE 0.086 Grade norm
RWSD 0.458 Accuracy
PARus 0.574 Accuracy
ruTiE 0.562 Accuracy
MultiQ 0.112 / 0.063 F1 Exact match
CheGeKa 0.011 / 0.005 F1 Exact match
ruModAr 0.358 Exact match
MaMuRAMu 0.456 Accuracy
ruMultiAr 0.64 Exact match
ruCodeEval 0.091 / 0.256 / 0.36 Pass@k
MathLogicQA 0.817 Accuracy
ruWorldTree 0.672 / 0.541 Accuracy F1 macro
ruOpenBookQA 0.543 / 0.439 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 0.602 Accuracy
ruMMLU 0.403 Accuracy
SimpleAr 0.816 Exact match
ruHumanEval 0.063 / 0.212 / 0.299 Pass@k
ruHHH 0.494
ruHateSpeech 0.506
ruDetox 0.071
ruEthics
Correct God Ethical
Virtue 0.042 0.193 0.211
Law 0.066 0.204 0.207
Moral 0.056 0.195 0.2
Justice 0.041 0.187 0.203
Utilitarianism 0.051 0.199 0.172

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:

Qwen3-0.6B

Model size

600.0M

Model type:

Opened

SFT

Additional links:

https://arxiv.org/abs/2505.09388

Architecture description:

Qwen3-0.6B is a decoder-only transformer language model from the Qwen3 family with approximately 0.6 billion parameters. The model is designed as a lightweight language model for general text understanding and generation tasks while retaining multilingual and reasoning capabilities.

Description of the training:

The model follows the same multi-stage training pipeline used across the Qwen3 series, including large-scale pretraining followed by post-training stages such as instruction tuning and alignment to improve instruction following and response quality.

Pretrain data:

The model was pretrained on a large multilingual corpus of approximately 36 trillion tokens covering 119 languages, including web documents, books, code repositories, and technical materials.

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
Qwen3-0.6B
MERA
0.1 0.067 0.4 0.133 0 0.1 0 - 0.033 0.033 0 0.033 0 0 0.033 0.233 0 0.067 0.033 0 0 0.033 0.067 0.033 0.1 0.158 0.1 0.167 0.1 0.133 0.067
Model, team Honest Helpful Harmless
Qwen3-0.6B
MERA
0.475 0.559 0.448
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
Qwen3-0.6B
MERA
0.326 0.313 0.421 0.585 0.389 0.458 0.466 0.383 0.364 0.372 0.237 0.333 0.33 0.454 0.421 0.431 0.42 0.403 0.422 0.435 0.283 0.304 0.39 0.416 0.304 0.47 0.287 0.389 0.298 0.36 0.52 0.529 0.411 0.455 0.426 0.5 0.5 0.432 0.377 0.429 0.414 0.364 0.483 0.79 0.457 0.454 0.436 0.622 0.312 0.344 0.33 0.468 0.413 0.408 0.6 0.479 0.435
Model, team SIM FL STA
Qwen3-0.6B
MERA
0.434 0.492 0.541
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
Qwen3-0.6B
MERA
0.356 0.465 0.367 0.463 0.434 0.397 0.362 0.386 0.288 0.508 0.654 0.508 0.208 0.419 0.287 0.333 0.505 0.422 0.579 0.456 0.281 0.424 0.644 0.467 0.622 0.515 0.346 0.281 0.667 0.444 0.711 0.385 0.411 0.509 0.394 0.625 0.511 0.533 0.544 0.508 0.416 0.429 0.556 0.889 0.603 0.844 0.379 0.818 0.446 0.579 0.511 0.391 0.62 0.455 0.256 0.327 0.5
Coorect
Good
Ethical
Model, team Virtue Law Moral Justice Utilitarianism
Qwen3-0.6B
MERA
0.042 0.066 0.056 0.041 0.051
Model, team Virtue Law Moral Justice Utilitarianism
Qwen3-0.6B
MERA
0.193 0.204 0.195 0.187 0.199
Model, team Virtue Law Moral Justice Utilitarianism
Qwen3-0.6B
MERA
0.211 0.207 0.2 0.203 0.172
Model, team Women Men LGBT Nationalities Migrants Other
Qwen3-0.6B
MERA
0.537 0.429 0.471 0.541 0.286 0.508