Missing prompt_tokens and total_tokens in Usage Object Returned in DeepSeek Model

[k99k5]

What Happened?

{
    "id": "74196500-8b9b-475a-aab3-381cec604587",
    "object": "chat.completion.chunk",
    "created": 1736352401,
    "choices": [
      {
        "index": 0,
        "finish_reason": "stop",
        "message": {
          "role": "assistant",
          "content": "\nif (isset($_POST['like'])) {\n    echo 'Thank you for your support!';\n}\n?>\n"
        }
      }
    ],
    "model": "deepseek-chat",
    "usage": {
      "completion_tokens": 25
    }
  }

What Should Have Happened?

The official API returns the following response:

{"id":"577145fa-34cb-4cc4-b937-6f8e03eafa79","object":"chat.completion.chunk","created":1736355607,"model":"deepseek-chat","system_fingerprint":"fp_3a5770e1b4","choices":[{"index":0,"delta":{"content":""},"logprobs":null,"finish_reason":"stop"}],"usage":{"prompt_tokens":9,"completion_tokens":11,"total_tokens":20,"prompt_cache_hit_tokens":0,"prompt_cache_miss_tokens":9}}

Relevant Code Snippet

No response

Your Twitter/LinkedIn

No response

[narengogi]

Thanks for reporting this @k99k5
This should be pretty straightforward to fix

[narengogi]

Hey @k99k5 I'm able to see the prompt_tokens in response for deepseek-chat model used through the gateway, can you provide a sample cURL that is not working for you?

request

curl --location 'https://api.portkey.ai/v1/chat/completions' \
--header 'x-portkey-provider: deepseek' \
--header 'Authorization: sk-' \
--header 'Content-Type: application/json' \
--header 'x-portkey-api-key: ' \
--data '{
        "model": "deepseek-chat",
        "messages": [
          {"role": "system", "content": "You are a helpful assistant."},
          {"role": "user", "content": "Tell me a story about love!"}
        ],
        "max_tokens": 100,
        "stream": false
}'

response

{
    "id": "85916956-abde-4b9a-99be-7d19a389a281",
    "object": "chat.completion",
    "created": 1737632471,
    "model": "deepseek-chat",
    "provider": "deepseek",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "Of course! Here's a story about love:\n\n---\n\n**The Garden of Two Hearts**\n\nOnce upon a time, in a quiet village nestled between rolling hills, there lived a young woman named Elara. She was known for her kindness and her love for flowers. Every morning, she would tend to her garden, where roses, lilies, and daisies bloomed in vibrant colors. The villagers often said her garden was the most beautiful in the land, but Elara felt something was missing"
            },
            "finish_reason": "length"
        }
    ],
    "usage": {
        "prompt_tokens": 16,
        "completion_tokens": 100,
        "total_tokens": 116
    }
}

[k99k5]

@narengogi curl --location 'https://api.portkey.ai/v1/chat/completions'
--header 'x-portkey-provider: deepseek'
--header 'x-portkey-config: xxxxxxxxxxx'
--header 'Content-Type: application/json'
--header 'x-portkey-api-key: xxxxxx'
--data '{
"model": "deepseek-chat",
"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Tell me a story about game!"}
],
"max_tokens": 100,
"stream": true,
"stream_options": {"include_usage": true}
}'

[narengogi]

added the fix for this in
#897