Datasetpublic

AgentPublic/evalap-comparing-openweight-with-openaigpt-oss-120b-113

Comparing openweight with openai/gpt-oss-120b (ID: 113) Comparing openweight Albert-API with specific judge openai/gpt-oss-120b Overview This dataset contains 24 experiments from the EvalAP evaluation platform. Datasets: Assistant IA - QA, MFS_questions_v01 Models evaluated: openweight-large, openweight-medium, openweight-small Metrics: energy_consumption, generation_time, gwp_consumption, judge_notator, judge_precision, nb_tokens_completion, nb_tokens_prompt… See the full description on the dataset page: https://huggingface.co/datasets/AgentPublic/evalap-comparing-openweight-with-openaigpt-oss-120b-113.

sourceHugging Faceupdated 8mo agoView on Hugging Face

0likes34downloads

Dataset Card

Comparing openweight with openai/gpt-oss-120b (ID: 113)

Comparing openweight Albert-API with specific judge openai/gpt-oss-120b

Overview

This dataset contains 24 experiments from the EvalAP evaluation platform.

Datasets: Assistant IA - QA, MFSquestionsv01

Models evaluated: openweight-large, openweight-medium, openweight-small

Metrics: energyconsumption, generationtime, gwpconsumption, judgenotator, judgeprecision, nbtokenscompletion, nbtokens_prompt

Scores

MFSquestionsv01

model	energy_consumption	generation_time	gwp_consumption	judge_notator	judge_precision	nb_tokens_completion	nb_tokens_prompt
openai/gpt-oss-120b Albert	0.02 ± 0.01	14.46 ± 4.56	0.00 ± 0.00	6.70 ± 3.07	0.26 ± 0.44	1747.21 ± 615.21	38.90 ± 9.26
Mistral small	0.00 ± 0.00	10.97 ± 5.62	0.00 ± 0.00	6.54 ± 2.84	0.16 ± 0.37	456.58 ± 165.10	38.90 ± 9.26
Qwen/Qwen3-VL-8B-Thinking	0.00 ± 0.00	8.53 ± 19.81	0.00 ± 0.00	6.96 ± 2.82	0.21 ± 0.40	1365.38 ± 613.20	38.90 ± 9.26

Assistant IA - QA

model	energy_consumption	generation_time	gwp_consumption	judge_notator	judge_precision	nb_tokens_completion	nb_tokens_prompt
openai/gpt-oss-120b Albert	0.01 ± 0.01	10.23 ± 4.94	0.00 ± 0.00	6.41 ± 3.36	0.24 ± 0.43	1194.73 ± 714.03	29.59 ± 7.05
Mistral small	0.00 ± 0.00	8.54 ± 4.92	0.00 ± 0.00	6.04 ± 3.20	0.11 ± 0.32	326.14 ± 139.16	29.59 ± 7.05
Qwen/Qwen3-VL-8B-Thinking	0.00 ± 0.00	4.90 ± 3.35	0.00 ± 0.00	5.58 ± 3.52	0.19 ± 0.39	1024.26 ± 692.94	29.59 ± 7.05

Usage

Use the dropdown above to select an experiment configuration.