Create a model

POST

model

curl --request POST \
  --url https://aiapi.scope3.com/model \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "id": "gpt-4-turbo",
  "name": "GPT-4 Turbo",
  "family": "gpt",
  "hugging_face_path": "EleutherAI/gpt-neo-2.7B",
  "benchmark_model_id": "gpt-4-turbo",
  "estimated_use_life_days": 365,
  "estimated_requests_per_day": 1000,
  "fine_tuned_from_model_id": "gpt-4-turbo",
  "aliases": [
    "claude-latest"
  ],
  "total_params_billions": 175,
  "number_of_experts": 7,
  "params_per_expert_billions": 8,
  "training_usage_energy_kwh": 1013.1,
  "training_usage_emissions_kgco2e": 1013.1,
  "training_usage_water_l": 1013.1,
  "training_embodied_emissions_kgco2e": 11013.1,
  "training_embodied_water_l": 11013.1
}'

{
  "id": "gpt-4-turbo",
  "aliases": [
    "claude-latest",
    "claude-3-sonnet-current"
  ],
  "name": "GPT-4 Turbo",
  "family": "gpt",
  "hugging_face_path": "EleutherAI/gpt-neo-2.7B",
  "benchmark_model_id": "GPTJ-6B",
  "total_params_billions": 175,
  "number_of_experts": 7,
  "params_per_expert_billions": 8,
  "tensor_parallelism": 1,
  "datatype": "fp8",
  "task": "text-generation",
  "training_usage_energy_kwh": 1013.1,
  "training_usage_emissions_kgco2e": 1013.1,
  "training_usage_water_l": 1013.1,
  "training_embodied_emissions_kgco2e": 11013.1,
  "training_embodied_water_l": 11013.1,
  "estimated_use_life_days": 1013.1,
  "estimated_requests_per_day": 1013.1,
  "fine_tuned_from_model_id": "llama_31_8b",
  "customer_id": 123,
  "created_at": "2023-11-07T05:31:56Z",
  "updated_at": "2023-11-07T05:31:56Z",
  "created_by": "<string>"
}

Authorizations

Authorization

string

header

required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

Create a new model Note on permissions:

Custom models are visible only to their owners
Global models are visible to all users
Admins can see and manage all models

string

required

Example:

"gpt-4-turbo"

number_of_experts

integer

required

Example:

7

name

string

Example:

"GPT-4 Turbo"

family

string

Example:

"gpt"

hugging_face_path

string

Example:

"EleutherAI/gpt-neo-2.7B"

benchmark_model_id

string

Example:

"gpt-4-turbo"

estimated_use_life_days

number

Example:

365

estimated_requests_per_day

number

Example:

1000

fine_tuned_from_model_id

string

Example:

"gpt-4-turbo"

aliases

string[]

Example:

["claude-latest"]

total_params_billions

number

Example:

175

params_per_expert_billions

number

Example:

8

training_usage_energy_kwh

number

Example:

1013.1

training_usage_emissions_kgco2e

number

Example:

1013.1

training_usage_water_l

number

Example:

1013.1

training_embodied_emissions_kgco2e

number

Example:

11013.1

training_embodied_water_l

number

Example:

11013.1

Response

201

application/json

Model created successfully

string

required

Example:

"gpt-4-turbo"

aliases

string[]

required

List of aliases for this model; must be globally-unique with id

Example:

["claude-latest", "claude-3-sonnet-current"]

number_of_experts

integer

required

Example:

7

customer_id

integer

required

ID of the customer who owns this node (visible to admins only)

created_at

string

required

updated_at

string

required

created_by

string

required

ID of the user who created the node (admin or owner only)

name

string

Example:

"GPT-4 Turbo"

family

string

Example:

"gpt"

hugging_face_path

string

Example:

"EleutherAI/gpt-neo-2.7B"

benchmark_model_id

string

Example:

"GPTJ-6B"

total_params_billions

number

Example:

175

params_per_expert_billions

number

Example:

8

tensor_parallelism

integer

Example:

1

datatype

enum<string>

Available options:

fp8,

fp8-e4m3,

fp8-e5m2,

fp16,

tf32,

fp32,

fp64,

bfloat8,

bfloat16,

bf16,

int4,

int8,

int16,

int32,

int64,

uint4,

uint8,

uint16,

uint32,

uint64

task

enum<string>

Common types of AI/ML models and their primary functions:

Text-based models for natural language processing
Vision models for image analysis and generation
Audio models for speech and sound processing
Multimodal models that combine different types of inputs/outputs
Specialized models for specific use cases

Available options:

text-generation,

chat,

text-embedding,

text-classification,

sentiment-analysis,

named-entity-recognition,

question-answering,

summarization,

translation,

image-classification,

object-detection,

image-segmentation,

image-generation,

image-to-text,

text-to-image,

style-transfer,

face-detection,

facial-recognition,

speech-to-text,

text-to-speech,

speaker-identification,

audio-classification,

music-generation,

multimodal-embedding,

multimodal-generation,

visual-question-answering,

recommendation-system,

reinforcement-learning,

anomaly-detection,

time-series-forecasting,

clustering

training_usage_energy_kwh

number

Example:

1013.1

training_usage_emissions_kgco2e

number

Example:

1013.1

training_usage_water_l

number

Example:

1013.1

training_embodied_emissions_kgco2e

number

Example:

11013.1

training_embodied_water_l

number

Example:

11013.1

estimated_use_life_days

number

Example:

1013.1

estimated_requests_per_day

number

Example:

1013.1

fine_tuned_from_model_id

string

Example:

"llama_31_8b"

List models Get a specific model (global or custom)

curl --request POST \
  --url https://aiapi.scope3.com/model \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "id": "gpt-4-turbo",
  "name": "GPT-4 Turbo",
  "family": "gpt",
  "hugging_face_path": "EleutherAI/gpt-neo-2.7B",
  "benchmark_model_id": "gpt-4-turbo",
  "estimated_use_life_days": 365,
  "estimated_requests_per_day": 1000,
  "fine_tuned_from_model_id": "gpt-4-turbo",
  "aliases": [
    "claude-latest"
  ],
  "total_params_billions": 175,
  "number_of_experts": 7,
  "params_per_expert_billions": 8,
  "training_usage_energy_kwh": 1013.1,
  "training_usage_emissions_kgco2e": 1013.1,
  "training_usage_water_l": 1013.1,
  "training_embodied_emissions_kgco2e": 11013.1,
  "training_embodied_water_l": 11013.1
}'

{
  "id": "gpt-4-turbo",
  "aliases": [
    "claude-latest",
    "claude-3-sonnet-current"
  ],
  "name": "GPT-4 Turbo",
  "family": "gpt",
  "hugging_face_path": "EleutherAI/gpt-neo-2.7B",
  "benchmark_model_id": "GPTJ-6B",
  "total_params_billions": 175,
  "number_of_experts": 7,
  "params_per_expert_billions": 8,
  "tensor_parallelism": 1,
  "datatype": "fp8",
  "task": "text-generation",
  "training_usage_energy_kwh": 1013.1,
  "training_usage_emissions_kgco2e": 1013.1,
  "training_usage_water_l": 1013.1,
  "training_embodied_emissions_kgco2e": 11013.1,
  "training_embodied_water_l": 11013.1,
  "estimated_use_life_days": 1013.1,
  "estimated_requests_per_day": 1013.1,
  "fine_tuned_from_model_id": "llama_31_8b",
  "customer_id": 123,
  "created_at": "2023-11-07T05:31:56Z",
  "updated_at": "2023-11-07T05:31:56Z",
  "created_by": "<string>"
}

Guides

Integrations

API Reference

Create a model

Authorizations

Body

Response