Frontier models through one simple API

Start with GPT 5.5, Kimi 2.6, GLM 5.1, and MiniMax 2.7. Compare speed and build without a subscription.

4
launch models
55tps
top throughput
1
API
0
subscriptions

Model infrastructure

Build with advanced AI models. Customize with your data. Launch with confidence.

Distributed AI compute network illustration

A globally distributed compute network engineered for resilient, high-performance delivery.

Secure cloud infrastructure illustration

Comprehensive security, governance, and reliability for mission-critical AI workloads.

Inference throughput meter illustration

Fast inference and efficient scaling for demanding production traffic.

Quality latency and cost deployment surface illustration

Smart deployment choices that balance quality, latency, and cost.

One endpoint

Use the same API shape across every launch model.

Clear speed

Throughput is visible before you choose a model.

Built for teams

Add policy, routing, and billing controls when the backend is ready.

Pricing

Flat-rate open-source models. Credits when you need more.

Subscribe for a generous daily token pool on open-source models, or top up credits for premium models and occasional usage.

Pay as you go

Customcredit top-ups

Top up credits when you need premium models, overflow capacity, or occasional API usage without a monthly plan.

Included usage

Credits for flexible usage

  • No monthly commitment
  • Use credits across metered models
  • OpenAI-compatible API access
  • Usage logs and balance tracking
Buy credits
Best start

Plus

$39/month

Flat-rate access for builders who want room for everyday coding, prototypes, and personal agents.

Included usage

75M open-source tokens / day

  • Daily allowance for open-source models
  • Credits still work for premium models
  • Built for coding sessions and agents
  • Everything in Pay as you go
Get Plus
More capacity

Pro

$120/month

More daily throughput for production agents, heavier workflows, and teams that route a lot of traffic.

Included usage

350M open-source tokens / day

  • Higher daily open-source allowance
  • Best fit for production traffic
  • Credits cover premium model usage
  • Everything in Plus
Get Pro