MODELAXY
INITIALIZING MODEL GALAXY / 2026
MODELAXY
Console

MODEL GALAXY / 模型星系

MODELAXY

One connection.
Every model in orbit.

Call commercial and local models through one stable interface. Modelaxy manages routing, recovery, usage, and privacy.

Start free Read API docs
API SERVICE AVAILABLEUSAGE PRICING / READY TO SCALE
SCROLL TO ORBIT

01 / STABLE INTERFACE

Models change.
Your interface doesn’t.

Every request enters one service layer for capability matching, availability switching, and usage records. Connect your product once and keep accessing a model ecosystem that never stops evolving.

  • One key for every enabled service profile
  • Automatic capacity failover when a provider degrades
  • Usage, latency, budgets, and spend visible per key
31+mainstream model families
9APIs and local runtimes
4stable service profiles
1usage and billing view

02 / SERVICE PROFILES

Four paths into orbit.

Choose the right capability for every product experience, from live interaction to deep analysis.

01 / 04

live interaction

Fast

tokenlab-fastFor chat, autocomplete, in-product assistants, and experiences that need immediate feedback.

03 / MODEL ECOSYSTEM

The Model Galaxy keeps expanding.

Connect leading commercial APIs, regional model services, and private open-source deployments. Customers keep using stable service names without seeing which underlying model handled a request.

COMPATIBLE APIS & RUNTIMES

OpenAI Chat CompletionsAnthropic MessagesGemini generateContentOpenAI-compatiblevLLMOllamaSGLangTGILM Studio

01GLOBAL COMMERCIAL

OpenAI GPTOpenAI o-seriesAnthropic ClaudeGoogle GeminixAI GrokMistralCodestral

02CHINA MODEL SERVICES

DeepSeekAlibaba QwenKimiZhipu GLMMiniMaxByteDance DoubaoTencent HunyuanBaidu ERNIEiFlytek SparkBaichuanStepFun01.AI Yi

03OPEN & PRIVATE

LlamaQwen OpenDeepSeek OpenMistral / MixtralGemmaPhiGLM OpenYi OpenInternLMLLaVAFalconDBRX

Model availability depends on workspace configuration, service region, and provider authorization.

04 / TRANSPARENT PRICING

Clear pricing. Pay for what runs.

No subscription required. Prepaid service credits settle against normalized units, with request-level usage and final cost visible in the console.

Add service credits
SERVICE PROFILEINPUT / 1M UNITSOUTPUT / 1M UNITSORCHESTRATION / REQUEST
Fasttokenlab-fast
INPUT / 1M UNITS$0.35OUTPUT / 1M UNITS$1.20ORCHESTRATION / REQUEST$0.00002
Generaltokenlab-general
INPUT / 1M UNITS$1.50OUTPUT / 1M UNITS$6.00ORCHESTRATION / REQUEST$0.00005
Reasoningtokenlab-reasoning
INPUT / 1M UNITS$4.00OUTPUT / 1M UNITS$16.00ORCHESTRATION / REQUEST$0.00030
Privatetokenlab-private
INPUT / 1M UNITS$2.00OUTPUT / 1M UNITS$8.00ORCHESTRATION / REQUEST$0.00010

Settled in USD. Final charges are recorded in the console ledger.

DATA POLICY / ACTIVE

05 / CONTROL & GOVERNANCE

Your data is never the price of routing.

tokenlab-private supports sensitive workloads. Teams can also choose zero-retention per request and set workspace-wide budgets and logging policies.

  • Request-level privacy modes
  • No prompt or response retention by default
  • Visible usage, latency, and cost

06 / GET STARTED

Put your next AI feature in orbit.

MODEL GALAXY

Open the Modelaxy console for free