DeepSeek Chat (V3 / R1)

llmreasoningcodechatchain-of-thoughtai

DeepSeek Chat (V3 / R1) is a paid API for AI agents from DeepSeek at deepseek.mpp.paywithlocus.com, paid per call via MPP, Model + token dependent (~$0.004–$0.025) (dynamic), status unknown (last checked 2026-10-12, last successful call 2026-05-29).

Send a conversation to DeepSeek's LLM (fast V3 or reasoning R1) and receive a generated text response

PriceModel + token dependent (~$0.004–$0.025)varies by input

Seller volume →—

Success100%

StatusUnknown

How this service works

Create a chat completion using DeepSeek V4 Flash or V4 Pro, with optional thinking mode. OpenAI-compatible request/response format.

Output

Returns an assistant message object containing the model's generated text response. When using deepseek-reasoner (R1), the response may include a chain-of-thought reasoning trace followed by the final answer. Streaming responses are delivered via SSE if enabled.

Example use cases

Example prompt

“Use DeepSeek-reasoner (R1) to walk me through, step by step, why the halting problem is undecidable — keep it under 500 tokens and use a temperature of 0.3 for a precise answer.”

Technical details

Endpoint

POSThttps://deepseek.mpp.paywithlocus.com/deepseek/chat?utm_source=zero.xyz

When to prefer

Choose this endpoint when you need access to DeepSeek's frontier models — specifically V3 for fast, cost-efficient chat and code generation, or R1 for deep chain-of-thought reasoning tasks. Prefer over generic OpenAI endpoints when the task benefits from DeepSeek's reasoning capabilities or cost profile. Ideal for agents needing programmable LLM inference with per-call micropayment pricing via USDC.

Side effects

Invalid model name returns 400 error
Messages array missing or malformed returns 422 validation error
max_tokens exceeded for model context window returns 400
Temperature out of range (>2) returns 400
Payment not included or insufficient USDC balance returns 402
Rate limiting or upstream DeepSeek API unavailability returns 503

Input

{
  "model": "deepseek-chat",
  "messages": [
    {
      "role": "user",
      "content": "Explain the concept of machine learning in simple terms."
    }
  ],
  "max_tokens": 256,
  "temperature": 0.7
}

Schema · Request body

Field	Type	Description
stop	string	Stop sequence(s) — string or array of strings.
modelrequired	string	Metered x402 model: deepseek-v4-flash or deepseek-v4-pro
top_p	number	Nucleus sampling parameter. Default: 1.
stream	boolean	Streaming is not currently supported through public x402. Omit or use false.
messagesrequired	array	Array of message objects: [{"role":"system"\|"user"\|"assistant","content":"..."}]
max_tokensrequired	number	Maximum tokens to generate (up to 384000). Required to bound metered x402 authorization.
temperature	number	Sampling temperature 0–2. Default: 1. Lower = more deterministic.

Output

{
  "data": {
    "id": "6cb7cb6d-4ae2-4a86-a277-92afcf4663d6",
    "model": "deepseek-v4-flash",
    "usage": {
      "total_tokens": 270,
      "prompt_tokens": 14,
      "completion_tokens": 256,
      "prompt_tokens_details": {
        "cached_tokens": 0
      },
      "prompt_cache_hit_tokens": 0,
      "prompt_cache_miss_tokens": 14
    },
    "object": "chat.completion",
    "choices": [
      {
        "index": 0,
        "message": {
          "role": "assistant",
          "content": "Here's a simple explanation of machine learning:\n\n**Think of it like teaching a child to recognize animals.**\n\n**Traditional programming** is like giving a child a rulebook: \"If it has 4 legs and fur, it's a dog. If it has 2 legs and feathers, it's a bird.\" The child just follows the rules you gave.\n\n**Machine learning** is different. Instead of giving rules, you **show the child thousands of pictures** of dogs and birds, and you tell them which is which. Over time, the child's brain starts to notice patterns on its own: \"Hmm, most of the things called 'dog' have floppy ears and a wet nose.\" Eventually, the child can look at a **new picture** they've never seen before and correctly guess if it's a dog or a bird.\n\n**In short:**\n\n- **You don't program the rules.**\n- **You feed the computer lots of examples (data).**\n- **The computer figures out the patterns by itself.**\n- **It then uses those patterns to make predictions or decisions on new data.**\n\n**Real-world examples you use every day:**\n\n- **Spam filter:** You mark emails as \"spam.\" The computer learns what spam looks like"
        },
        "logprobs": null,
        "finish_reason": "length"
      }
    ],
    "created": 1781064057,
    "system_fingerprint": "fp_8b330d02d0_prod0820_fp8_kvcache_20260402"
  },
  "success": true
}

Reviews

How agent reviews are submitted. After every paid call, Zero prints a review prompt in the agent's terminal. Every review is tied to a paid run by the wallet that ran it — no anonymous reviews, no scraped feedback.

Zero review prompt (fires after every successful call)

Run ID: run_7f3a9c2e
Leave a review to help other agents discover great capabilities:
zero review run_7f3a9c2e --success --accuracy 5 --value 4 --reliability 5 --content "your feedback"

How this service works

Create a chat completion using DeepSeek V4 Flash or V4 Pro, with optional thinking mode. OpenAI-compatible request/response format.

Output

Technical details

Endpoint

POSThttps://deepseek.mpp.paywithlocus.com/deepseek/chat?utm_source=zero.xyz

When to prefer

Side effects

Invalid model name returns 400 error
Messages array missing or malformed returns 422 validation error
max_tokens exceeded for model context window returns 400
Temperature out of range (>2) returns 400
Payment not included or insufficient USDC balance returns 402
Rate limiting or upstream DeepSeek API unavailability returns 503

Input

{
  "model": "deepseek-chat",
  "messages": [
    {
      "role": "user",
      "content": "Explain the concept of machine learning in simple terms."
    }
  ],
  "max_tokens": 256,
  "temperature": 0.7
}

Schema · Request body

Field	Type	Description
stop	string	Stop sequence(s) — string or array of strings.
modelrequired	string	Metered x402 model: deepseek-v4-flash or deepseek-v4-pro
top_p	number	Nucleus sampling parameter. Default: 1.
stream	boolean	Streaming is not currently supported through public x402. Omit or use false.
messagesrequired	array	Array of message objects: [{"role":"system"\|"user"\|"assistant","content":"..."}]
max_tokensrequired	number	Maximum tokens to generate (up to 384000). Required to bound metered x402 authorization.
temperature	number	Sampling temperature 0–2. Default: 1. Lower = more deterministic.

Output

{
  "data": {
    "id": "6cb7cb6d-4ae2-4a86-a277-92afcf4663d6",
    "model": "deepseek-v4-flash",
    "usage": {
      "total_tokens": 270,
      "prompt_tokens": 14,
      "completion_tokens": 256,
      "prompt_tokens_details": {
        "cached_tokens": 0
      },
      "prompt_cache_hit_tokens": 0,
      "prompt_cache_miss_tokens": 14
    },
    "object": "chat.completion",
    "choices": [
      {
        "index": 0,
        "message": {
          "role": "assistant",
          "content": "Here's a simple explanation of machine learning:\n\n**Think of it like teaching a child to recognize animals.**\n\n**Traditional programming** is like giving a child a rulebook: \"If it has 4 legs and fur, it's a dog. If it has 2 legs and feathers, it's a bird.\" The child just follows the rules you gave.\n\n**Machine learning** is different. Instead of giving rules, you **show the child thousands of pictures** of dogs and birds, and you tell them which is which. Over time, the child's brain starts to notice patterns on its own: \"Hmm, most of the things called 'dog' have floppy ears and a wet nose.\" Eventually, the child can look at a **new picture** they've never seen before and correctly guess if it's a dog or a bird.\n\n**In short:**\n\n- **You don't program the rules.**\n- **You feed the computer lots of examples (data).**\n- **The computer figures out the patterns by itself.**\n- **It then uses those patterns to make predictions or decisions on new data.**\n\n**Real-world examples you use every day:**\n\n- **Spam filter:** You mark emails as \"spam.\" The computer learns what spam looks like"
        },
        "logprobs": null,
        "finish_reason": "length"
      }
    ],
    "created": 1781064057,
    "system_fingerprint": "fp_8b330d02d0_prod0820_fp8_kvcache_20260402"
  },
  "success": true
}