AI Compute · Flat-Rate Monthly PlanBuilt for AI Agents

Flat-Rate AI Compute:
Choose Premium, Standard Unlimited

Running OpenClaw, Hermes Agent, or Cursor all day? Token bills add up fast.

TokenTable is a flat-rate monthly plan for AI compute — choose a Main model quota, Side models unlimited. No per-token billing. From $19 (NT$600)/month.

300+
Models Supported
∞
Unlimited Sides
99.9%
Uptime
$19
Starting from (NT$600)
Used by AI Agent builders
OpenClawOpenClaw·Hermes AgentHermes Agent·CursorCursor·WindsurfWindsurf·ClineCline·Claude CodeClaude Code·CodexCodex·Gemini CLIGemini CLI·AutoGenAutoGen·DifyDify·LobeChatLobeChat·Open WebUIOpen WebUI·Cherry StudioCherry Studio·AiderAider
How It Works

Like a restaurant set menu

Pick one Main course quota, unlimited Sides. TokenTable works the same — choose your Main model allowance; Side models run unlimited.

Main · Top-tier Models

Top models, per-plan quota

Claude Fable 5, Claude Fable 5.1, Claude Haiku 5.5, Qwen Flash, Qwen Plus

Complex reasoning, high-quality output — fixed monthly allowance by plan.

Starter · $0New-user kickoff: up to 20 flagship Main calls in your first 72 hours
Solo · $19 (NT$600)All-you-can-eat Sides — 30M priority tokens/mo
Pro · $49 (NT$1,590)All-you-can-eat Sides — 100M priority tokens/mo
Team · $129 (NT$4,190)All-you-can-eat Sides — 250M priority tokens/mo (team-shared)
Advanced · $199 (NT$6,390)All-you-can-eat Sides — 500M priority tokens/mo (team-shared)
Enterprise · Unlimited Sides
Sides · Unlimited

Budget models, all-you-can-eat

Smart routing auto-selects from a dynamic model pool — daily tasks, agent batch jobs, data processing. Run thousands of calls a day at no extra cost.

∞
Unlimited
Paid plans · monthly priority allowance
AI Router handles switching automatically
70% of tasks work fine on Sides. Our AI Router picks the right model by task type — only tapping your Main quota when deep reasoning is truly needed.
Predictable bills, never surprise charges
$19/$49/$129/$199 (NT$600 / 1,590 / 4,190 / 6,390) flat monthly. Sides run freely without extra charges — your CFO will love this.
One API key for all your agents
OpenClaw, Hermes, Cursor, Cline, AutoGen — just swap the base URL. No need for separate OpenAI/Anthropic/Google accounts.
AI Advisor in your language
Stuck on setup? The AI Advisor walks you through connecting Cursor, OpenClaw, Hermes — no English documentation needed.
ONE KEY · ALL CAPABILITIES

Your AI agent does more than chat

With one TokenTable API key, your agent can do all of this at once — no separate sign-ups, no separate bills.

💬
Text & reasoning
Chat, write, analyze, drive agent decisions — top models like Claude, GPT, and Gemini, all included.
👁️
See & read images
Hand it an image: understand content, OCR text, read screenshots and tables.
📄
Long context
Read a whole contract or codebase in one go (up to 1M tokens) without losing the thread.
💻
Write code
Write code, debug, generate docs — plug straight into Cursor / Cline / Claude Code.
🎨
Generate images
Text-to-image in one line, OpenAI Images API compatible — your agent can generate art too.
🎬
Generate video
Text-to-video in one line, ready to wire into automated pipelines.
🎙️
Generate voice
Text-to-speech in one line, multiple voices — narration for short video, audiobooks.
AI Advisor

Don't know how to set up? Just ask.

How do I run Claude Code on TokenTable? How to connect Cursor or Cline? Does the router switch between Main and Sides automatically? — Just ask.

✓Full support, no English docs required
✓All major AI Agent platforms supported
✓Any OpenAI-compatible tool works
AI AdvisorLive
3 tries left
Hi! I'm the TokenTable AI Advisor. Whether you use Claude Code, Cursor, Cline, Codex, or Dify — ask me anything about Token costs or how to connect.
Supported Models

300+ Models, AI Router selects automatically

Main models can be named explicitly. Sides are split into capability pools — image gen, video, coding — AI Router picks the best fit automatically.

Main · Top-tier Models (Named Selection)Per-plan quota
Model
Context
Specialty
claude-fable-5
Claude
1MMost capable (selectable)
claude-fable-5-1
Claude
1MMost capable (latest, selectable)
claude-haiku-5-5
Claude
1MNewest Haiku (fast, selectable)
claude-opus-4-8
Claude
1MFlagship reasoning
claude-opus-5
Claude
1MLatest flagship reasoning (selectable)
claude-opus-5-5
Claude
1MNewest Opus (selectable)
claude-sonnet-4-6
Claude
1MFast + intelligent
claude-sonnet-5
Claude
1MSonnet flagship (selectable)
claude-sonnet-5-5
Claude
1MNewest Sonnet (selectable)
gpt-5.5
GPT
1MGeneral purpose
gpt-5.6
GPT
1MGPT flagship — Sol (selectable)
gpt-5.6-luna
GPT
1MGPT flagship — Luna (cost-efficient, selectable)
gpt-5.6-terra
GPT
1MGPT flagship — Terra (balanced, selectable)
gpt-6-astra
GPT
1MGPT-6 flagship (selectable)
gpt-6-luna
GPT
1MGPT-6 — Luna (cost-efficient, selectable)
gpt-6-sol
GPT
1MGPT-6 — Sol (selectable)
gpt-6.1-sol
GPT
1MGPT-6.1 — Sol (selectable)
gpt-6.1-terra
GPT
1MGPT-6.1 — Terra (balanced, selectable)
grok-4.20-0309-non-reasoning
Grok
1MGrok next-gen (selectable)
grok-4.20-0309-reasoning
Grok
1MGrok next-gen reasoning (selectable)
grok-4.3
Grok
256kReal-time web
grok-4.5
Grok
256kGrok reasoning (selectable)
grok-4.6
Grok
500kGrok flagship (new)
grok-4.7
Grok
500kGrok latest (selectable)
gemini-3.5-flash
Gemini
1MGemini previous gen (selectable)
gemini-3.6-flash
Gemini
1MGemini flagship
gemini-3.8-flash
Gemini
1MGemini latest (selectable)
gemini-4-argon
Gemini
1MGemini 4 flagship reasoning (selectable)
mistral-large-4
Mistral
512kMistral Large 4 (EU, reasoning, preview)
mistral-large-latest
Mistral
256kMistral Large (EU)
qwen-flash
Qwen
1MQwen fast (low-latency)
qwen-plus
Qwen
1MQwen classic flagship
qwen3.5-plus
Qwen
1MQwen3.5 Plus vision-language (selectable)
qwen3.6-plus
Qwen
1MCN flagship
qwen3.7-max
Qwen
1MCN agent flagship (new)
qwen3.7-plus
Qwen
1MCN vision flagship (new)
qwen3.8-max
Qwen
1MCN flagship (new, vision)
deepseek-ai/DeepSeek-V4.1-Flash
DeepSeek
1MDeepSeek V4.1 Flash (selectable)
deepseek-v4-pro
DeepSeek
128kDeep reasoning Pro
deepseek-v4.1-flash
DeepSeek
1MDeep reasoning Flash 4.1
doubao-seed-2.0-lite
Doubao Seed 2.0 Lite
256kDoubao Seed 2.0 Lite (selectable)
doubao-seed-2.0-mini
Doubao Seed 2.0 Mini
256kDoubao Seed 2.0 Mini (selectable)
doubao-seed-2.1-lite
Doubao Seed 2.1 Lite
1MDoubao Seed 2.1 Lite (selectable)
doubao-seed-2.1-pro
Doubao Seed 2.1 Pro
1MDoubao Seed 2.1 Pro (selectable)
doubao-seed-2.1-turbo
Doubao Seed 2.1 Turbo
256kDoubao Seed 2.1 Turbo (selectable)
doubao-seed-character
Doubao Seed Character
128kDoubao Seed Character (role-play, selectable)
doubao-seed-evolving
Doubao Seed Evolving
1MDoubao Seed Evolving (latest coding/agent line, weekly-updated, selectable)
glm-5.1
GLM
1MZhipu GLM (selectable)
glm-5.2
GLM
1MZhipu GLM flagship (new)
glm-5.3-prime
GLM
1MZhipu GLM-5.3 Prime (fast, selectable)
kimi-k2.5
Kimi
256kKimi K2.5 coding (selectable)
kimi-k2.6
Kimi
128kCoding flagship
kimi-k2.7-code
Kimi
256kKimi K2.7 coding (new)
kimi/kimi-k2.8-preview
Kimi
1MKimi K2.8 preview (selectable)
kimi/kimi-k3
Kimi
1MKimi K3 flagship (selectable)
kimi/kimi-k3.1-pro
Kimi
1MKimi K3.1 Pro reasoning flagship (selectable)
kimi/kimi-k3.1-pro-highspeed
Kimi
1MKimi K3.1 Pro high-speed (selectable)
meta-models/Muse-Spark-Open-816B-A42B
Muse Spark Open 816B A42B
1MMeta Muse Spark Open 816B (selectable)
MiniMax-M2.5
MiniMax
1MCN text (selectable)
MiniMax/MiniMax-M2.7
MiniMax
1MCN agentic (MiniMax M2.7)
MiniMax/MiniMax-M3
MiniMax
1MCN flagship (MiniMax M3, MSA)
MiniMaxAI/MiniMax-M3.1-Flash
MiniMax
1MMiniMax M3.1 Flash (selectable)
nvidia/NVIDIA-Nemotron-3.5-Super-VL
NVIDIA Nemotron 3.5 Super VL
262kNVIDIA Nemotron 3.5 Super VL (selectable)
papaya-1.0-spicy
Papaya 1.0 Spicy
32kPapaya 1.0 Spicy (adult creative writing and prompts, selectable). Inputs: images as a public https URL or uploaded directly as base64 (image_url); .md/.markdown/.txt files uploaded as base64 (file part, up to 300 KB); MP4 video by public https URL only (video_url, up to 50 MB and 3 minutes, stream: true required; pictures only, no audio). Whole request up to 8 MiB
reflection/Beam-1.0-501B-A23B
Beam 1.0.501B A23B
1MReflection Beam 1.0 (selectable)
unisound/unisound-u2
Unisound U2
160kUnisound U2 agent flagship (selectable)
xiaomi/mimo-v2.6-pro-ultraspeed
Mimo V2.6 Pro Ultraspeed
1MXiaomi MiMo V2.6 Pro UltraSpeed (selectable)
zai-org/GLM-5.3-Flash
GLM
1MZhipu GLM Flash (selectable)
ZHIPU/GLM-5.3
GLM
1MZhipu GLM next-gen (selectable)
Sides · Capability Pools (AI Router Auto-selects)All-you-can-eat on paid plans · 300+ budget models
💬Chat / Writing
80+

Daily chat, writing, summarization

Pool examples:
Qwen3.7 FlashQwen FlashQwen Plus...and more
⚡Coding / Debug
40+

Code gen, bug fix, code review

Pool examples:
DeepSeek V4.1 FlashQwen3 Coder Flash...and more
👁️Vision
50+

OCR, image understanding, screenshot Q&A

Pool examples:
Qwen3 VL PlusQwen VL OCR...and more
🎨Image Generation
25+

Text-to-image: illustrations, avatars

Pool examples:
Wan2.7 ImageKling V3 Omni Image GenerationQwen Image 3.0Doubao Seedream 5.0 Pro...and more
🎬Video Generation
12+

Text/image to video, animation

Pool examples:
Wan2.6 T2vKling V3 Omni Video GenerationHappyhorse 1.1 T2vDoubao Seedance 2.5...and more
🎙️Voice Generation
4+

Text-to-speech, multiple voices for video narration

Pool examples:
Cosyvoice V2...and more
📐Long Context
15+

1M+ token context, book summarization, RAG

Pool examples:
Qwen Long...and more
🌍Translation
18+

100+ languages, more natural than Google Translate

Pool examples:
Qwen3.7 FlashQwen Flash...and more
🧠Reasoning

Complex logic, multi-step reasoning, deep analysis

Pool examples:
Qwen3.5 Flash...and more
📄Productivity

Document processing, report writing, office productivity

Pool examples:
Qwen PlusGlm 5.1...and more

Just tell your agent what you want to do — AI Router picks the right model from the pool automatically.

See the full main & side model list →

Compare

TokenTable vs Per-Token Billing

FeaturePer-Token PlatformTokenTable
Pricing ModelPer-token + 5.5% platform fee✅ Flat monthly, from $19 (NT$600)
Bill Predictability❌ Surprise charges anytime✅ Fixed monthly, from $19 (NT$600)
Model Selection❌ Research and pick yourself✅ AI Router auto-selects
Side models (budget tier)❌ Still billed per token✅ Truly unlimited
Main models (top tier)Per token✅ Quota-based, auto-fallback
Language / DocsEnglish only✅ Multilingual + AI Advisor
Best forEngineers who control every request✅ Teams focused on building agents
STABLE · COMPLIANT · HERE TO STAY

Why TokenTable, not a free relay?

There are plenty of "free unlimited Claude / GPT" relays out there — but that "free" is paid for with stability, compliance, and your data security. Roughly 60% of them vanish within months, taking your prepaid balance with them. For real work and your company's code, you need a platform that's still here next year.

🛡️
We're here, so is your balance
Legitimate channels and real payments — not shared or cracked accounts. We won't get banned and vanish overnight with your prepaid balance.
♾️
A legit "free unlimited"
Our 300+ side models are genuinely unlimited and free for light tasks — all-you-can-eat built on compliance, not on reverse-engineered subscriptions.
🔒
Your code stays private
Enterprises can switch on Confidential Mode (no content logging) plus a compliance attestation — safe to feed your company repo, no shady relay in the middle.
🧾
Invoiced, predictable monthly
Clean for company expensing with a locked monthly fee — no bans, no surprise bills to scare the CFO.
Get Started

Set up in 30 seconds, cut next month's bill in half

1
Sign up free, try it in your browser
30 seconds, no credit card. API Keys unlock with Solo ($19/mo) and up
2
Ask the AI Advisor how to connect
Tell it you use OpenClaw / Hermes / Cursor — it guides you in your language
3
Swap the base URL
Enter the URL below and you're done tokentable.asia/v1
4
Let your agent run, bill cut in half
AI Router picks the model — you do nothing
Pricing

Choose your plan

Paid plans include all-you-can-eat Sides plus a generous monthly priority allowance. Start free, upgrade or cancel anytime.

Starter
$0
✓New-user kickoff: up to 20 flagship Main calls in your first 72 hours
✓Then 5 calls/day: 2 mains + 3 sides (resets daily)
✓AI Advisor (basic)
Solo
$19/mo
NT$600
✓All-you-can-eat Sides — 30M priority tokens/mo
✓~10 Main/day
✓AI Advisor (full)
✓Top-Up Pack — refill anytime
ProBest Value
$49/mo
NT$1,590
✓All-you-can-eat Sides — 100M priority tokens/mo
✓~35 Main/day
✓AI Advisor (full)
✓Priority response
✓Top-Up Pack — refill anytime
TeamFor Teams
$129/mo
NT$4,190
✓All-you-can-eat Sides — 250M priority tokens/mo (team-shared)
✓AI Advisor (full)
✓Top-Up Pack — refill anytime
Advanced
$199/mo
NT$6,390
✓All-you-can-eat Sides — 500M priority tokens/mo (team-shared)
✓AI Advisor (full)
✓SLA 99.9%
✓Dedicated support
Enterprise
✓Unlimited Sides
✓Unlimited Main
✓Custom seat count
✓Dedicated infra
✓Custom SLA + contract

When Main quota runs out, AI Router auto-switches to Sides — no errors, no interruption.

FAQ

Common Questions

Get Started

Let your AI Agent run unlimited

Sides (AI Router smart selection) are unlimited on paid plans — new users get 20 flagship Main calls in their first 72 hours, then 5 calls/day after that. The more your agent runs, the more you save.

No credit card·No contract·Multilingual interface