Qwen2.5-Omni-7B

MERA Created at 20.11.2025 16:52

Ratings for leaderboard tasks

The table will scroll to the left

Board Result Attempted Score Coverage Place in the rating
Multi 0.302 0.302 1 3
Images 0.211 0.211 1 12
Audio 0.464 0.464 1 3
Video 0.42 0.42 1 8

Tasks

The table will scroll to the left

Task Modality Result Metric
0.432
EM JudgeScore
0.273
EM F1
0.598
EM JudgeScore
0.186
EM JudgeScore
0.21
EM JudgeScore
0.602
EM JudgeScore
0.13
EM JudgeScore
0.055
EM JudgeScore
0.511
EM JudgeScore
0.359
EM JudgeScore
0.16
EM JudgeScore
0.28
EM JudgeScore
0.382
EM JudgeScore
0.367
EM JudgeScore
0.467
EM JudgeScore
0.094
EM JudgeScore
culture 0.057 / 0.112
business 0.076 / 0.151
medicine 0.059 / 0.114
social_sciences 0.092 / 0.175
fundamental_sciences 0.056 / 0.101
applied_sciences 0.074 / 0.155
0.174
EM JudgeScore
biology 0.136 / 0.253
chemistry 0.109 / 0.22
physics 0.18 / 0.3
economics 0.108 / 0.179
ru 0.077 / 0.143
all 0.132 / 0.229
0.153
EM JudgeScore
biology 0.053 / 0.123
chemistry 0.045 / 0.179
physics 0.101 / 0.237
science 0.171 / 0.293

Information about the submission

Mera version
1.0.0
Torch Version
2.8.0
The version of the codebase
7e640aa
CUDA version
12.6
Precision of the model weights
bfloat16
Seed
1234
Batch
1
Transformers version
4.57.1
The number of GPUs and their type
1 x A100
Architecture
openai-chat-completions

Team:

MERA

Name of the ML model:

Qwen2.5-Omni-7B

Model size

7.0B

Model type:

Opened

SFT

Inference parameters

The size of the context:
32768