Create training session

curl --request POST \
  --url https://api.together.ai/v1/rl/training-sessions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model_resources_id": "123e4567-e89b-12d3-a456-426614174000",
  "resume_from_checkpoint_id": "123e4567-e89b-12d3-a456-426614174000",
  "lora_config": {
    "rank": 32,
    "alpha": 64,
    "dropout": 0,
    "seed": "59",
    "train_unembed": true
  },
  "display_name": "gsm8k-experiment-2",
  "metadata": {
    "wandb": {
      "entity": "example-org",
      "project": "grpo-gsm8k",
      "group": "gsm8k-35b-sweep",
      "run_name": "exp2-thinking-4k-ctx",
      "run_id": "abc123",
      "url": "https://wandb.ai/example-org/example-project/runs/run-id"
    }
  },
  "load_optimizer": true
}
'

import requests

url = "https://api.together.ai/v1/rl/training-sessions"

payload = {
    "model_resources_id": "123e4567-e89b-12d3-a456-426614174000",
    "resume_from_checkpoint_id": "123e4567-e89b-12d3-a456-426614174000",
    "lora_config": {
        "rank": 32,
        "alpha": 64,
        "dropout": 0,
        "seed": "59",
        "train_unembed": True
    },
    "display_name": "gsm8k-experiment-2",
    "metadata": { "wandb": {
            "entity": "example-org",
            "project": "grpo-gsm8k",
            "group": "gsm8k-35b-sweep",
            "run_name": "exp2-thinking-4k-ctx",
            "run_id": "abc123",
            "url": "https://wandb.ai/example-org/example-project/runs/run-id"
        } },
    "load_optimizer": True
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    model_resources_id: '123e4567-e89b-12d3-a456-426614174000',
    resume_from_checkpoint_id: '123e4567-e89b-12d3-a456-426614174000',
    lora_config: {rank: 32, alpha: 64, dropout: 0, seed: '59', train_unembed: true},
    display_name: 'gsm8k-experiment-2',
    metadata: {
      wandb: {
        entity: 'example-org',
        project: 'grpo-gsm8k',
        group: 'gsm8k-35b-sweep',
        run_name: 'exp2-thinking-4k-ctx',
        run_id: 'abc123',
        url: 'https://wandb.ai/example-org/example-project/runs/run-id'
      }
    },
    load_optimizer: true
  })
};

fetch('https://api.together.ai/v1/rl/training-sessions', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.together.ai/v1/rl/training-sessions",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'model_resources_id' => '123e4567-e89b-12d3-a456-426614174000',
    'resume_from_checkpoint_id' => '123e4567-e89b-12d3-a456-426614174000',
    'lora_config' => [
        'rank' => 32,
        'alpha' => 64,
        'dropout' => 0,
        'seed' => '59',
        'train_unembed' => true
    ],
    'display_name' => 'gsm8k-experiment-2',
    'metadata' => [
        'wandb' => [
                'entity' => 'example-org',
                'project' => 'grpo-gsm8k',
                'group' => 'gsm8k-35b-sweep',
                'run_name' => 'exp2-thinking-4k-ctx',
                'run_id' => 'abc123',
                'url' => 'https://wandb.ai/example-org/example-project/runs/run-id'
        ]
    ],
    'load_optimizer' => true
  ]),
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>",
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.together.ai/v1/rl/training-sessions"

	payload := strings.NewReader("{\n  \"model_resources_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"resume_from_checkpoint_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"lora_config\": {\n    \"rank\": 32,\n    \"alpha\": 64,\n    \"dropout\": 0,\n    \"seed\": \"59\",\n    \"train_unembed\": true\n  },\n  \"display_name\": \"gsm8k-experiment-2\",\n  \"metadata\": {\n    \"wandb\": {\n      \"entity\": \"example-org\",\n      \"project\": \"grpo-gsm8k\",\n      \"group\": \"gsm8k-35b-sweep\",\n      \"run_name\": \"exp2-thinking-4k-ctx\",\n      \"run_id\": \"abc123\",\n      \"url\": \"https://wandb.ai/example-org/example-project/runs/run-id\"\n    }\n  },\n  \"load_optimizer\": true\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.together.ai/v1/rl/training-sessions")
  .header("Authorization", "Bearer <token>")
  .header("Content-Type", "application/json")
  .body("{\n  \"model_resources_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"resume_from_checkpoint_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"lora_config\": {\n    \"rank\": 32,\n    \"alpha\": 64,\n    \"dropout\": 0,\n    \"seed\": \"59\",\n    \"train_unembed\": true\n  },\n  \"display_name\": \"gsm8k-experiment-2\",\n  \"metadata\": {\n    \"wandb\": {\n      \"entity\": \"example-org\",\n      \"project\": \"grpo-gsm8k\",\n      \"group\": \"gsm8k-35b-sweep\",\n      \"run_name\": \"exp2-thinking-4k-ctx\",\n      \"run_id\": \"abc123\",\n      \"url\": \"https://wandb.ai/example-org/example-project/runs/run-id\"\n    }\n  },\n  \"load_optimizer\": true\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.together.ai/v1/rl/training-sessions")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"model_resources_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"resume_from_checkpoint_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"lora_config\": {\n    \"rank\": 32,\n    \"alpha\": 64,\n    \"dropout\": 0,\n    \"seed\": \"59\",\n    \"train_unembed\": true\n  },\n  \"display_name\": \"gsm8k-experiment-2\",\n  \"metadata\": {\n    \"wandb\": {\n      \"entity\": \"example-org\",\n      \"project\": \"grpo-gsm8k\",\n      \"group\": \"gsm8k-35b-sweep\",\n      \"run_name\": \"exp2-thinking-4k-ctx\",\n      \"run_id\": \"abc123\",\n      \"url\": \"https://wandb.ai/example-org/example-project/runs/run-id\"\n    }\n  },\n  \"load_optimizer\": true\n}"

response = http.request(request)
puts response.read_body

{
  "id": "123e4567-e89b-12d3-a456-426614174000",
  "metadata": {
    "wandb": {
      "entity": "example-org",
      "project": "grpo-gsm8k",
      "group": "gsm8k-35b-sweep",
      "run_name": "exp2-thinking-4k-ctx",
      "run_id": "abc123",
      "url": "https://wandb.ai/example-org/example-project/runs/run-id"
    }
  },
  "status": "TRAINING_SESSION_STATUS_RUNNING",
  "inference_checkpoints": [
    {
      "id": "123e4567-e89b-12d3-a456-426614174000",
      "step": "42",
      "created_at": "2026-01-02T00:00:00Z",
      "registration": {
        "model": {
          "id": "ml_abc123",
          "revision_id": "rv_def456"
        },
        "adapter": {
          "id": "ml_abc123",
          "revision_id": "rv_def456"
        }
      }
    }
  ],
  "training_checkpoints": [
    {
      "id": "123e4567-e89b-12d3-a456-426614174000",
      "step": "42",
      "created_at": "2026-01-02T00:00:00Z",
      "registration": {
        "id": "ml_abc123",
        "revision_id": "rv_def456"
      }
    }
  ],
  "step": "100",
  "created_at": "2026-01-02T00:00:00Z",
  "updated_at": "2026-01-02T00:00:05Z",
  "model_resources_id": "123e4567-e89b-12d3-a456-426614174000",
  "base_model": "Qwen/Qwen3.5-4B",
  "created_by": "user_123",
  "policy_state": {
    "trainer_step": "7",
    "target_weights_version": "7",
    "applied_weights_version": "6",
    "pending_publish": true
  },
  "display_name": "gsm8k-experiment-2",
  "error": {
    "code": "TRAINING_SESSION_ERROR_CODE_RESOURCE_AT_CAPACITY",
    "message": "The session could not start because the selected resource has no available capacity.",
    "occurred_at": "2026-01-02T00:00:05Z"
  },
  "resume_from_checkpoint_id": "123e4567-e89b-12d3-a456-426614174000",
  "lora_config": {
    "rank": 32,
    "alpha": 64,
    "dropout": 0,
    "seed": "59",
    "train_unembed": true
  }
}

{
  "error": {
    "message": "<string>",
    "type": "<string>",
    "param": null,
    "code": null
  }
}

POST

training-sessions

Create training session

curl --request POST \
  --url https://api.together.ai/v1/rl/training-sessions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model_resources_id": "123e4567-e89b-12d3-a456-426614174000",
  "resume_from_checkpoint_id": "123e4567-e89b-12d3-a456-426614174000",
  "lora_config": {
    "rank": 32,
    "alpha": 64,
    "dropout": 0,
    "seed": "59",
    "train_unembed": true
  },
  "display_name": "gsm8k-experiment-2",
  "metadata": {
    "wandb": {
      "entity": "example-org",
      "project": "grpo-gsm8k",
      "group": "gsm8k-35b-sweep",
      "run_name": "exp2-thinking-4k-ctx",
      "run_id": "abc123",
      "url": "https://wandb.ai/example-org/example-project/runs/run-id"
    }
  },
  "load_optimizer": true
}
'

import requests

url = "https://api.together.ai/v1/rl/training-sessions"

payload = {
    "model_resources_id": "123e4567-e89b-12d3-a456-426614174000",
    "resume_from_checkpoint_id": "123e4567-e89b-12d3-a456-426614174000",
    "lora_config": {
        "rank": 32,
        "alpha": 64,
        "dropout": 0,
        "seed": "59",
        "train_unembed": True
    },
    "display_name": "gsm8k-experiment-2",
    "metadata": { "wandb": {
            "entity": "example-org",
            "project": "grpo-gsm8k",
            "group": "gsm8k-35b-sweep",
            "run_name": "exp2-thinking-4k-ctx",
            "run_id": "abc123",
            "url": "https://wandb.ai/example-org/example-project/runs/run-id"
        } },
    "load_optimizer": True
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    model_resources_id: '123e4567-e89b-12d3-a456-426614174000',
    resume_from_checkpoint_id: '123e4567-e89b-12d3-a456-426614174000',
    lora_config: {rank: 32, alpha: 64, dropout: 0, seed: '59', train_unembed: true},
    display_name: 'gsm8k-experiment-2',
    metadata: {
      wandb: {
        entity: 'example-org',
        project: 'grpo-gsm8k',
        group: 'gsm8k-35b-sweep',
        run_name: 'exp2-thinking-4k-ctx',
        run_id: 'abc123',
        url: 'https://wandb.ai/example-org/example-project/runs/run-id'
      }
    },
    load_optimizer: true
  })
};

fetch('https://api.together.ai/v1/rl/training-sessions', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.together.ai/v1/rl/training-sessions",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'model_resources_id' => '123e4567-e89b-12d3-a456-426614174000',
    'resume_from_checkpoint_id' => '123e4567-e89b-12d3-a456-426614174000',
    'lora_config' => [
        'rank' => 32,
        'alpha' => 64,
        'dropout' => 0,
        'seed' => '59',
        'train_unembed' => true
    ],
    'display_name' => 'gsm8k-experiment-2',
    'metadata' => [
        'wandb' => [
                'entity' => 'example-org',
                'project' => 'grpo-gsm8k',
                'group' => 'gsm8k-35b-sweep',
                'run_name' => 'exp2-thinking-4k-ctx',
                'run_id' => 'abc123',
                'url' => 'https://wandb.ai/example-org/example-project/runs/run-id'
        ]
    ],
    'load_optimizer' => true
  ]),
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>",
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.together.ai/v1/rl/training-sessions"

	payload := strings.NewReader("{\n  \"model_resources_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"resume_from_checkpoint_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"lora_config\": {\n    \"rank\": 32,\n    \"alpha\": 64,\n    \"dropout\": 0,\n    \"seed\": \"59\",\n    \"train_unembed\": true\n  },\n  \"display_name\": \"gsm8k-experiment-2\",\n  \"metadata\": {\n    \"wandb\": {\n      \"entity\": \"example-org\",\n      \"project\": \"grpo-gsm8k\",\n      \"group\": \"gsm8k-35b-sweep\",\n      \"run_name\": \"exp2-thinking-4k-ctx\",\n      \"run_id\": \"abc123\",\n      \"url\": \"https://wandb.ai/example-org/example-project/runs/run-id\"\n    }\n  },\n  \"load_optimizer\": true\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.together.ai/v1/rl/training-sessions")
  .header("Authorization", "Bearer <token>")
  .header("Content-Type", "application/json")
  .body("{\n  \"model_resources_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"resume_from_checkpoint_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"lora_config\": {\n    \"rank\": 32,\n    \"alpha\": 64,\n    \"dropout\": 0,\n    \"seed\": \"59\",\n    \"train_unembed\": true\n  },\n  \"display_name\": \"gsm8k-experiment-2\",\n  \"metadata\": {\n    \"wandb\": {\n      \"entity\": \"example-org\",\n      \"project\": \"grpo-gsm8k\",\n      \"group\": \"gsm8k-35b-sweep\",\n      \"run_name\": \"exp2-thinking-4k-ctx\",\n      \"run_id\": \"abc123\",\n      \"url\": \"https://wandb.ai/example-org/example-project/runs/run-id\"\n    }\n  },\n  \"load_optimizer\": true\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.together.ai/v1/rl/training-sessions")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Authorization"] = 'Bearer <token>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"model_resources_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"resume_from_checkpoint_id\": \"123e4567-e89b-12d3-a456-426614174000\",\n  \"lora_config\": {\n    \"rank\": 32,\n    \"alpha\": 64,\n    \"dropout\": 0,\n    \"seed\": \"59\",\n    \"train_unembed\": true\n  },\n  \"display_name\": \"gsm8k-experiment-2\",\n  \"metadata\": {\n    \"wandb\": {\n      \"entity\": \"example-org\",\n      \"project\": \"grpo-gsm8k\",\n      \"group\": \"gsm8k-35b-sweep\",\n      \"run_name\": \"exp2-thinking-4k-ctx\",\n      \"run_id\": \"abc123\",\n      \"url\": \"https://wandb.ai/example-org/example-project/runs/run-id\"\n    }\n  },\n  \"load_optimizer\": true\n}"

response = http.request(request)
puts response.read_body

{
  "id": "123e4567-e89b-12d3-a456-426614174000",
  "metadata": {
    "wandb": {
      "entity": "example-org",
      "project": "grpo-gsm8k",
      "group": "gsm8k-35b-sweep",
      "run_name": "exp2-thinking-4k-ctx",
      "run_id": "abc123",
      "url": "https://wandb.ai/example-org/example-project/runs/run-id"
    }
  },
  "status": "TRAINING_SESSION_STATUS_RUNNING",
  "inference_checkpoints": [
    {
      "id": "123e4567-e89b-12d3-a456-426614174000",
      "step": "42",
      "created_at": "2026-01-02T00:00:00Z",
      "registration": {
        "model": {
          "id": "ml_abc123",
          "revision_id": "rv_def456"
        },
        "adapter": {
          "id": "ml_abc123",
          "revision_id": "rv_def456"
        }
      }
    }
  ],
  "training_checkpoints": [
    {
      "id": "123e4567-e89b-12d3-a456-426614174000",
      "step": "42",
      "created_at": "2026-01-02T00:00:00Z",
      "registration": {
        "id": "ml_abc123",
        "revision_id": "rv_def456"
      }
    }
  ],
  "step": "100",
  "created_at": "2026-01-02T00:00:00Z",
  "updated_at": "2026-01-02T00:00:05Z",
  "model_resources_id": "123e4567-e89b-12d3-a456-426614174000",
  "base_model": "Qwen/Qwen3.5-4B",
  "created_by": "user_123",
  "policy_state": {
    "trainer_step": "7",
    "target_weights_version": "7",
    "applied_weights_version": "6",
    "pending_publish": true
  },
  "display_name": "gsm8k-experiment-2",
  "error": {
    "code": "TRAINING_SESSION_ERROR_CODE_RESOURCE_AT_CAPACITY",
    "message": "The session could not start because the selected resource has no available capacity.",
    "occurred_at": "2026-01-02T00:00:05Z"
  },
  "resume_from_checkpoint_id": "123e4567-e89b-12d3-a456-426614174000",
  "lora_config": {
    "rank": 32,
    "alpha": 64,
    "dropout": 0,
    "seed": "59",
    "train_unembed": true
  }
}

{
  "error": {
    "message": "<string>",
    "type": "<string>",
    "param": null,
    "code": null
  }
}

Authorizations

Authorization

string

header

default:default

required

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

Body

application/json

model_resources_id

string

required

ID of the model resource to use for this training session.

Example:

"123e4567-e89b-12d3-a456-426614174000"

resume_from_checkpoint_id

string

Checkpoint ID to resume from. LoRA training checkpoints may resume on another model resource with compatible base-model weights. Full-weight training checkpoints require the original base model.

Example:

"123e4567-e89b-12d3-a456-426614174000"

lora_config

object

LoRA adapter configuration for the session

Show child attributes

display_name

string

Optional display name used to identify the training session

Maximum string length: 128

Example:

"gsm8k-experiment-2"

metadata

object

Optional auxiliary metadata to associate with the training session

Show child attributes

load_optimizer

boolean

Whether to restore optimizer state and step from a training checkpoint. Omitted or true restores them; false loads weights only with a fresh optimizer and step 0. Not valid for inference checkpoints, which have no optimizer state.

Example:

true

Response

Training session details

A training session and its current state

string

required

ID of the training session

Example:

"123e4567-e89b-12d3-a456-426614174000"

metadata

object

required

Auxiliary metadata associated with the training session

Show child attributes

status

enum<string>

default:TRAINING_SESSION_STATUS_UNSPECIFIED

required

Status of the training session

Available options:

TRAINING_SESSION_STATUS_UNSPECIFIED,

TRAINING_SESSION_STATUS_CREATING,

TRAINING_SESSION_STATUS_RUNNING,

TRAINING_SESSION_STATUS_STOPPED,

TRAINING_SESSION_STATUS_STOPPING,

TRAINING_SESSION_STATUS_ERROR,

TRAINING_SESSION_STATUS_EXPIRED

Example:

"TRAINING_SESSION_STATUS_RUNNING"

inference_checkpoints

object[]

required

List of saved inference checkpoints for this session

Show child attributes

training_checkpoints

object[]

required

List of saved training checkpoints for this session

Show child attributes

step

default:0

required

Current training step

Example:

"100"

created_at

string<date-time>

required

Timestamp when the training session was created

Example:

"2026-01-02T00:00:00Z"

updated_at

string<date-time>

required

Timestamp when the training session was last updated

Example:

"2026-01-02T00:00:05Z"

model_resources_id

string

required

ID of the model resource used by this training session.

Example:

"123e4567-e89b-12d3-a456-426614174000"

base_model

string

required

Base model the session trains, taken from the model resource it is attached to

Example:

"Qwen/Qwen3.5-4B"

created_by

string

required

ID of the user who created the training session

Example:

"user_123"

policy_state

object

required

Session-scoped policy and weight versions for this session

Show child attributes

display_name

string

Display name used to identify the training session

Maximum string length: 128

Example:

"gsm8k-experiment-2"

error

object

Structured detail for the training session's current error. Set when the session is in an error state.

Show child attributes

resume_from_checkpoint_id

string

Checkpoint ID this session was resumed from

Example:

"123e4567-e89b-12d3-a456-426614174000"

lora_config

object

LoRA adapter configuration. Present only for sessions running on a LoRA-enabled model resource.

Show child attributes

⌘I

TOGETHER APIS

SDK LIBRARIES

ERROR CODES

Create training session

Authorizations

Body

Response