πŸš€ Now in open beta

Premium Chinese LLMs.
Global pricing.

Access DeepSeek V4, Qwen3, GLM-5 and more β€” all with a single OpenAI-compatible API.
No Chinese phone number required. Pay in USD. Global support.

View Pricing
⚑ One-line migration from OpenAI β€” just change your base URL

COST COMPARISON (PER 1M TOKENS)

$30.00
OpenAI GPT-5.5
β†’
$3.00
TokenAgent (Qwen3)
β†’
$1.00
TokenAgent (DeepSeek Flash)

Simple, transparent pricing

Pay only for what you use. No monthly commitments. No hidden fees.

BEST VALUE

DeepSeek V4 Flash

Speed + quality balance, 1M context

$0.50 / 1M input
  • βœ“ Output: $1.00 / 1M tokens
  • βœ“ Cache hit: $0.01 / 1M tokens
  • βœ“ 1M context window
  • βœ“ 384K max output
  • βœ“ Reasoning & non-reasoning modes
POPULAR
MOST CAPABLE

Qwen3-235B

Alibaba's flagship, near GPT-5 level

$1.00 / 1M input
  • βœ“ Output: $3.00 / 1M tokens
  • βœ“ 128K context window
  • βœ“ Multilingual (δΈ­ζ–‡/English)
  • βœ“ Code & reasoning specialist
  • βœ“ OpenAI-compatible
PREMIUM

DeepSeek V4 Pro

Anthropic API compatible, maximum power

$3.00 / 1M input
  • βœ“ Output: $6.00 / 1M tokens
  • βœ“ Cache hit: $0.01 / 1M tokens
  • βœ“ 1M context window
  • βœ“ Native Anthropic API format
  • βœ“ 75% off promo until May 31

One-line migration from OpenAI

If you already use OpenAI SDK, switching takes 5 seconds.

Python 
from openai import OpenAI

# Before: client = OpenAI(api_key="sk-openai-...")
# After:  Just change the base URL
client = OpenAI(
    api_key="your-token-agent-key",
    base_url="https://openrouter.ai/api/v1"  # ← That's it
)

response = client.chat.completions.create(
    model="deepseek/deepseek-v4-flash",
    messages=[{"role": "user", "content": "Hello!"}]
)
Node.js 
import OpenAI from 'openai';

const client = new OpenAI({
    apiKey: "your-token-agent-key",
    baseURL: "https://openrouter.ai/api/v1"
});

const response = await client.chat.completions.create({
    model: "deepseek/deepseek-v4-flash",
    messages: [{ role: "user", content: "Hello!" }]
});

Available models

All models are available via a single API endpoint.

Model Input ($/1M) Output ($/1M) Cache Hit Context
DeepSeek V4 Flash
Best value, 384K output		 $0.50 $1.00 $0.01 1M
DeepSeek V4 Pro
Anthropic API compatible		 $3.00 $6.00 $0.01 1M
Qwen3-235B
Alibaba flagship, ~GPT-5 level		 $1.00 $3.00 β€” 128K
GLM-5
Zhipu AI's latest		 $1.50 $4.00 β€” 128K
More models coming soon
Minimax, Baichuan, Yi, and others		 β€” β€” β€” β€”

Ready to save 90% on AI costs?

Join developers who are already using Chinese LLMs through TokenAgent. Get your API key in under 2 minutes.

Get Your Free API Key β†’

No Chinese phone number required Β· Pay in USD Β· 24/7 support