Datasets:

o-xyz
/

ori_benchmark_leaderboard_metrics

Modalities:

Tabular

Text

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Viewer

Auto-converted to Parquet Duplicate

Split (1)

train · 77 rows

Model_id int64 12 42	Provider stringlengths 3 15	Source stringclasses 1 value	Model stringlengths 5 32	MMLU-Pro float64 0.67 0.82	GPQA float64 0.43 0.67	HLE float64 0.04 0.05	LiveCodeBench float64 0.14 0.41	SciCode float64 0.17 0.36	HumanEval float64 0.82 0.93	Math-500 float64 0.7 0.94	AIME float64 0.1 0.52	Speed(Tokens/Sec) float64 16.2 2.74k	Cost float64 0.07 8
12	CentML	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	16.2	0.8
12	Deepinfra	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	26.1	0.52
12	Fireworks AI	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	258.1	0.9
12	Hyperbolic	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	25.6	1.25
12	Microsoft Azure	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	77.8	2
12	Nebius	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	21.9	0.75
12	Novita	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	24.5	0.57
12	SambaNova	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	264.6	1.13
12	kluster	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	21.8	1.25
12	together	Open	DeepSeek V3 (Mar' 25)	0.82	0.66	0.05	0.41	0.36	0.92	0.94	0.52	93.9	1.25
14	CentML	Open	Llama 4 Maverick (FP8)	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	126.9	0.2
14	Deepinfra	Open	Llama 4 Maverick (FP8)	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	115.2	0.3
14	Lambda	Open	Llama 4 Maverick (FP8)	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	119.5	0.28
14	Microsoft Azure	Open	Llama 4 Maverick (FP8)	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	56.4	0.61
14	Novita	Open	Llama 4 Maverick (FP8)	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	83	0.34
14	kluster	Open	Llama 4 Maverick (FP8)	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	163.3	0.35
14	together	Open	Llama 4 Maverick (FP8)	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	127.3	0.41
15	AWS	Open	Llama 4 Maverick	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	299.8	0.42
15	Fireworks AI	Open	Llama 4 Maverick	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	179.8	0.39
15	SambaNova	Open	Llama 4 Maverick	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	794.2	0.92
15	groq	Open	Llama 4 Maverick	0.81	0.67	0.05	0.4	0.33	0.88	0.89	0.39	277	0.3
21	Alibaba	Open	Qwen2.5 Max	0.76	0.59	0.05	0.36	0.34	0.93	0.84	0.23	51.1	2.8
37	MiniMax	Open	MiniMax-Text-01	0.76	0.58	0.04	0.25	0.25	0.86	0.75	0.13	32.5	0.42
20	Deepinfra	Open	DeepSeek V3 (Dec '24)	0.75	0.56	0.04	0.36	0.35	0.91	0.89	0.25	34.8	0.59
20	Fireworks AI	Open	DeepSeek V3 (Dec '24)	0.75	0.56	0.04	0.36	0.35	0.91	0.89	0.25	71	1.31
20	Microsoft Azure	Open	DeepSeek V3 (Dec '24)	0.75	0.56	0.04	0.36	0.35	0.91	0.89	0.25	69.3	2
20	Nebius	Open	DeepSeek V3 (Dec '24)	0.75	0.56	0.04	0.36	0.35	0.91	0.89	0.25	23.4	0.75
20	Novita	Open	DeepSeek V3 (Dec '24)	0.75	0.56	0.04	0.36	0.35	0.91	0.89	0.25	30.7	0.89
23	AWS	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	162.1	0.29
23	CentML	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	114.4	0.1
23	Cerebras	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	2,737.5	0.7
23	Deepinfra	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	47.8	0.15
23	Fireworks AI	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	164.9	0.26
23	Lambda	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	119.7	0.14
23	Microsoft Azure	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	36.3	0.34
23	Novita	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	59.2	0.2
23	SambaNova	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	794.7	0.47
23	groq	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	569.4	0.17
23	kluster	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	91.8	0.71
23	together	Open	Llama 4 Scout	0.75	0.59	0.04	0.3	0.17	0.83	0.84	0.28	120	0.28
30	Lambda	Open	Llama 3.1 405B (FP8)	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	33	0.8
31	Databricks	Open	Llama 3.1 405B	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	38.4	7.5
31	Deepinfra	Open	Llama 3.1 405B	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	23.6	0.9
31	Fireworks AI	Open	Llama 3.1 405B	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	94.6	3
31	Hyperbolic	Open	Llama 3.1 405B	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	98.7	4
31	Microsoft Azure	Open	Llama 3.1 405B	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	31.6	8
31	SambaNova	Open	Llama 3.1 405B	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	171.1	6.25
32	AWS	Open	Llama 3.1 405B Latency Optimized	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	86.2	3
33	Nebius	Open	Llama 3.1 405B Base	0.73	0.52	0.04	0.31	0.3	0.85	0.7	0.21	33	1.5
34	Alibaba	Open	Qwen2.5 72B	0.72	0.49	0.04	0.28	0.27	0.88	0.86	0.16	51.6	0.2
34	Deepinfra	Open	Qwen2.5 72B	0.72	0.49	0.04	0.28	0.27	0.88	0.86	0.16	33.7	0.27
34	Fireworks AI	Open	Qwen2.5 72B	0.72	0.49	0.04	0.28	0.27	0.88	0.86	0.16	77.2	0.9
34	Hyperbolic	Open	Qwen2.5 72B	0.72	0.49	0.04	0.28	0.27	0.88	0.86	0.16	20.1	0.4
34	Nebius	Open	Qwen2.5 72B	0.72	0.49	0.04	0.28	0.27	0.88	0.86	0.16	24.9	0.2
35	Nebius	Open	Qwen2.5 72B Fast	0.72	0.49	0.04	0.28	0.27	0.88	0.86	0.16	70.5	0.38
36	together	Open	Qwen2.5 72B Turbo	0.72	0.49	0.04	0.28	0.27	0.88	0.86	0.16	87.3	1.2
39	Cohere	Open	Command A	0.71	0.53	0.05	0.29	0.28	0.82	0.82	0.1	90.4	4.38
25	Lambda	Open	Llama 3.3 70B (FP8)	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	57.3	0.17
26	AWS	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	261.3	0.71
26	CentML	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	145.9	0.5
26	Cerebras	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	2,566.9	0.94
26	Deepinfra	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	31.1	0.27
26	Fireworks AI	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	114.5	0.9
26	FriendliAI	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	168	0.6
26	Hyperbolic	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	34.4	0.4
26	Microsoft Azure	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	52.6	0.71
26	Novita	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	108.4	0.2
26	SambaNova	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	456.1	0.75
26	groq	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	408.1	0.64
26	kluster	Open	Llama 3.3 70B	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	31.9	0.7
27	Nebius	Open	Llama 3.3 70B Fast	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	139.8	0.38
28	Nebius	Open	Llama 3.3 70B Base	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	40	0.2
29	Deepinfra	Open	Llama 3.3 70B (Turbo, FP8)	0.71	0.5	0.04	0.29	0.26	0.86	0.77	0.3	31.7	0.2
38	Deepinfra	Open	Phi-4	0.71	0.57	0.04	0.23	0.26	0.87	0.81	0.14	41.3	0.09
38	Microsoft Azure	Open	Phi-4	0.71	0.57	0.04	0.23	0.26	0.87	0.81	0.14	31.5	0.22
38	Nebius	Open	Phi-4	0.71	0.57	0.04	0.23	0.26	0.87	0.81	0.14	116.2	0.15
42	Deepinfra	Open	Gemma 3 27B	0.67	0.43	0.05	0.14	0.21	0.89	0.88	0.25	37.6	0.07

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

ORI (Ocean Recommendation Integration) Dataset

Overview

This dataset contains consolidated model metrics used by the ORI (Ocean Recommendation Integration) system for AI model recommendations. The data includes performance metrics across various benchmarks, speed measurements, and cost information for different AI models and providers.

Structure

The dataset contains the following key information:

Model_id: Unique identifier for the model
Provider: The provider of the model (e.g., Nebius, Hyperbolic, etc.)
Source: Source of the benchmark data
Model: Name of the model
Benchmark columns: Various benchmark scores for each model
Speed(Tokens/Sec): Processing speed in tokens per second
Cost: Cost metric for the model

Usage

This dataset is used by the ORI system to recommend appropriate AI models based on:

Query similarity to benchmark tasks
User preferences for accuracy, speed, and cost
Provider availability

Citation

If you use this dataset in your research or applications, please cite:

@dataset{ori_dataset,
  author = {OXYZ AI Team},
  title = {ORI Model Metrics Dataset},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/o-xyz/ori_dataset}}
}

License

Downloads last month: 4

Number of rows:

Total file size:

13.2 kB