TokenRouter Free API Guide: 2 Permanent Free Models + 133 Aggregated, One-Click OpenAI/Claude/Gemini Compatible Gateway

Updated: 2026-08-29 · Official: https://www.tokenrouter.com · Models: 133 · Free verified: 2026-08-29 live pricing check

TL;DR: TokenRouter by PaleBlueDot AI is a unified AI model hub that normalizes leading LLMs into OpenAI, Claude and Gemini-compatible APIs. One API key + one base URL for 133 models, with personal quick-start and enterprise-grade governance.

Why TokenRouter

• One key for 133 models: Text, image, video, audio across Qwen, GLM, DeepSeek, Gemini, Claude, GPT, Grok, Seedream, Kling, MiniMax and more
• Triple-compatible endpoints: openai / openai-response / anthropic / gemini / video-generation etc., auto-matched per model
• Permanent free models: qwen/qwen3.8-max-free and nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free at $0.00 / 1M tokens input & output (Pay as you go, verified on model pages)
• Dynamic Global Routing + Multi-Channel Failover: Global nodes pick the optimal path in real time; auto-failover across upstream providers and owned inference cloud
• Smart Caching + Real-Time Cost Governance: Request filtering & result reuse to cut tokens; pre/in/post-request controls to prevent runaway spend
• Enterprise controls: Centralized billing, quota by member/team/department, audit-ready logs, zero data retention (only model/timestamp/tokens/cost logged, never content)

Free API in Detail (Verified)

See: https://www.tokenrouter.com/models/qwen/qwen3.8-max-free/ and https://www.tokenrouter.com/models/nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free/ — both show $0.0000 per 1M. Ideal for zero-cost integration tests, prompt tuning and demos.

Paid reference on the same hub:

• qwen/qwen3.8-max: $1.00 / $3.00 per 1M
• deepseek/deepseek-v4-pro: $1.32 / $3.96 per 1M
• Compare on https://www.tokenrouter.com/models

5-Minute Quick Start

1) Sign up & create a key

1. Go to https://www.tokenrouter.com → Sign Up (email verification)
2. Console → API Keys → Create Key (e.g., quickstart)
3. (Optional) Providers → Add provider key (OpenAI/Anthropic etc.) for specific routing; skip if you only use the free models
4. Models → copy model id (e.g., qwen/qwen3.8-max-free)
5. Chat to test online; Usage Logs & Balance to verify spend

2) Call via OpenAI-compatible API

curl https://api.tokenrouter.com/v1/chat/completions \
  -H "Authorization: Bearer $TOKENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3.8-max-free",
    "messages": [{"role": "user", "content": "Introduce TokenRouter in one sentence"}]
  }'

from openai import OpenAI
client = OpenAI(api_key="YOUR_TOKENROUTER_API_KEY", base_url="https://api.tokenrouter.com/v1")
resp = client.chat.completions.create(
    model="nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free",
    messages=[{"role": "user", "content": "Explain TokenRouter in one sentence."}],
)
print(resp.choices[0].message.content)

import OpenAI from "openai";
const client = new OpenAI({ apiKey: process.env.TOKENROUTER_API_KEY, baseURL: "https://api.tokenrouter.com/v1" });
const res = await client.chat.completions.create({
  model: "qwen/qwen3.8-max-free",
  messages: [{ role: "user", content: "Hello from TokenRouter!" }],
});
console.log(res.choices[0].message.content);

Migration tip: Existing OpenAI projects usually only need to change baseURL + apiKey; use anthropic / gemini formats where the model page indicates them.

3) Tips for free models

• Validate pipeline & prompts on :free models, then switch to flagship paid models for quality/cost comparison
• Video/image models (wan3.0-video, seedream-5.0-pro, kling-3.0-turbo etc.) use video-generation / image-generation endpoints
• Use Usage Logs for model / tokens / cost / latency and enable Smart Caching for repeated prompts

Personal vs Enterprise

Enterprise adds member/role management, weekly/monthly quotas, org Overview dashboard and full-chain traceability.

Pricing & Cost Optimization

• Transparent pricing on every model card (input/output/cache) before integration
• Routing modes: cost / quality / speed / specific model
• Owned inference cloud for open models via self-hosted GPU clusters
• Zero data retention: only metadata logged; content never retained

FAQ

Q: Are free models rate-limited? Check the console/model page for live limits. Free tier is ideal for dev & light traffic; use paid models or enterprise quotas for high concurrency.

Q: How much code to change from OpenAI? Typically just baseURL + apiKey + model id (qwen/qwen3.8-max-free etc.); messages/temperature stay OpenAI-compatible.

Q: 401 / No provider keys? Verify Authorization: Bearer is your TokenRouter key, add provider keys if the route requires them, and confirm base URL matches console.

Official Resources

• Home: https://www.tokenrouter.com/
• Models: https://www.tokenrouter.com/models
• Free models: https://www.tokenrouter.com/models/qwen/qwen3.8-max-free/ · https://www.tokenrouter.com/models/nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free/
• Feature guide: https://www.tokenrouter.com/docs/tokenrouter-feature-guide/
• Getting started: https://www.tokenrouter.com/blog/how-do-i-get-started-with-tokenrouter-api/ · https://www.tokenrouter.com/blog/how-do-i-sign-up-for-tokenrouter-and-get-api-access/

🚀 Get Started: One-Click Free API Access

Want to call all the free models above with a single API key, no need to sign up for each provider? Apishare.cc provides a unified API Key — one key, 100+ models, free models at zero cost.

👉 Register on Apishare.cc → Get your unified API Key

📊 Want to see more free model rankings? Check out the Sep 2026 Free LLM API Rankings →

About the Free API Aggregator

The models covered in this guide are all served through the APIShare free API aggregator, which gives you one key for the whole catalog.

• Full model catalog: APIShare free API directory
• Sign up for a free trial key: Register and claim your API key