GigaChat 2 Pro

GIGACHAT Created at 12.03.2025 11:33

0.649

The overall result

Place in the rating

In the top by tasks:

ruWorldTree

The result on the task is higher than human

The task is one of the main ones

ruHumanEval

USE

The task is one of the main ones

MathLogicQA

The task is one of the main ones

ruModAr

The task is one of the main ones

ruCodeEval

The task is one of the main ones

Hide

Weak tasks:

RWSD

PARus

RCB

199

ruEthics

MultiQ

ruOpenBookQA

102

CheGeKa

ruMMLU

297

ruHateSpeech

287

ruDetox

ruHHH

ruMultiAr

SimpleAr

105

LCS

210

BPS

MaMuRAMu

+12

Hide

Ratings for leaderboard tasks

The table will scroll to the left

Task name	Result	Metric
LCS	0.138	Accuracy
RCB	0.562 / 0.392	Accuracy F1 macro
USE	0.534	Grade norm
RWSD	0.665	Accuracy
PARus	0.934	Accuracy
ruTiE	0.883	Accuracy
MultiQ	0.492 / 0.37	F1 Exact match
CheGeKa	0.296 / 0.248	F1 Exact match
ruModAr	0.943	Exact match
MaMuRAMu	0.831	Accuracy
ruMultiAr	0.396	Exact match
ruCodeEval	0.599 / 0.633 / 0.64	Pass@k
MathLogicQA	0.775	Accuracy
ruWorldTree	0.99 / 0.99	Accuracy F1 macro
ruOpenBookQA	0.93 / 0.746	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

0.938

Accuracy

ruMMLU

0.745

Accuracy

SimpleAr

0.994

Exact match

ruHumanEval

0.627 / 0.672 / 0.683

Pass@k

ruHHH

0.837

ruHateSpeech

0.634

ruDetox

0.142

ruEthics

	Correct	God	Ethical
Virtue	0.273	0.283	0.355
Law	0.274	0.276	0.34
Moral	0.297	0.297	0.376
Justice	0.236	0.249	0.311
Utilitarianism	0.229	0.251	0.29

Information about the submission:

Mera version

v.1.2.0

Torch Version

2.5.1

The version of the codebase

30667dc

CUDA version

12.4

Precision of the model weights

Seed

1234

Butch

Transformers version

4.49.0

The number of GPUs and their type

Architecture

gigachat_llms

Team:

GIGACHAT

Name of the ML model:

GigaChat 2 Pro

Link to the ML model:

https://developers.sber.ru/docs/ru/gigachat/models

Model type:

Closed

API

SFT

Architecture description:

GigaChat 2 Pro is a new Large Language Model (LLM) by Sber. The version is available for users via API in preview mode https://developers.sber.ru/portal/products/gigachat-api

Description of the training:

Pretrain data:

License:

Proprietary model by Sber

Inference parameters

Generation Parameters:
rucodeeval - do_sample=true;temperature=0.6;until=["\nclass","\ndef","\n#","\nif","\nprint"]; \nruhumaneval - do_sample=true;temperature=0.6;until=["\nclass","\ndef","\n#","\nif","\nprint"];

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:
API Default

Expand information

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
GigaChat 2 Pro GIGACHAT	0.833	0.6	0.833	0.3	0.433	0.767	0.4	-	0.367	0.167	0.167	0.2	0.633	0.267	0.1	0.683	0.267	0.3	0.567	0.167	0.6	0.733	0.6	0.433	0.367	0.767	0.7	0.6	0.8	0.7	0.833

Model, team	Honest	Helpful	Harmless
GigaChat 2 Pro GIGACHAT	0.803	0.814	0.897

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
GigaChat 2 Pro GIGACHAT	0.711	0.584	0.901	0.846	0.837	0.861	0.816	0.749	0.87	0.74	0.64	0.603	0.55	0.759	0.842	0.746	0.8	0.896	0.644	0.794	0.647	0.848	0.58	0.728	0.679	0.85	0.491	0.667	0.759	0.58	0.8	0.843	0.736	0.919	0.781	0.876	0.66	0.91	0.715	0.773	0.889	0.783	0.738	0.854	0.886	0.722	0.819	0.607	0.539	0.738	0.77	0.793	0.854	0.92	0.91	0.794	0.896

Model, team	SIM	FL	STA
GigaChat 2 Pro GIGACHAT	0.236	0.725	0.767

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
GigaChat 2 Pro GIGACHAT	0.689	0.901	0.8	0.713	0.895	0.845	0.672	0.737	0.865	0.785	0.808	0.8	0.492	0.845	0.789	0.778	0.785	0.844	0.825	0.86	0.86	0.949	0.933	0.905	0.822	0.909	0.885	0.754	0.912	0.844	0.844	0.91	0.786	0.86	0.758	0.857	0.956	0.844	0.789	0.785	0.878	0.905	0.867	0.978	0.931	0.933	0.897	0.932	0.892	0.93	0.844	0.855	0.823	0.766	0.721	0.83	0.844

Coorect

Good

Ethical

Model, team	Virtue	Law	Moral	Justice	Utilitarianism
GigaChat 2 Pro GIGACHAT	0.273	0.274	0.297	0.236	0.229

Model, team	Virtue	Law	Moral	Justice	Utilitarianism
GigaChat 2 Pro GIGACHAT	0.283	0.276	0.297	0.249	0.251

Model, team	Virtue	Law	Moral	Justice	Utilitarianism
GigaChat 2 Pro GIGACHAT	0.355	0.34	0.376	0.311	0.29

Model, team	Women	Men	LGBT	Nationalities	Migrants	Other
GigaChat 2 Pro GIGACHAT	0.843	0.6	0	0.135	0.286	0.803

GigaChat 2 Pro

Ratings for leaderboard tasks

Evaluation on open tasks:

Information about the submission:

Team:

Name of the ML model:

Link to the ML model:

Model type:

Architecture description:

Description of the training:

Pretrain data:

License:

Inference parameters

Ratings by subcategory

Confirm the deletion of the sub