EDGE NODE AI
Tokens as a Service
Connect to private LLM servers in your region.
Ultra-low latency inference.
Why Edge Node
Built for production inference
Ultra-Low Latency
Expect single-digit millisecond overhead on top of fast model speed.
Secure Access Keys
Scoped API keys provisioned instantly on subscription. Rotate or revoke keys at any time from your dashboard.
Grow Instantly
Start small and scale instantly to enterprise performance — all with zero downtime.
How It Works
From zero to inference in minutes
- 1
Subscribe to a model
Pick a model and plan. Your subscription activates immediately — no sales call required.
- 2
Get your key instantly
A scoped API key is provisioned on the spot, tied to your plan and usage limits.
- 3
Point your SDK at the gateway
Swap the base URL in any OpenAI- or Anthropic-compatible SDK and start sending requests.
$ export ANTHROPIC_BASE_URL="https://qwen3-coder.edge-node.ai/v1"
$ export ANTHROPIC_API_KEY="hwl_live_..."
$ claude
✓ Connected to Edge Node · Dallas, TX · 4ms
Models
The newest open models, day one
Frontier open-weight models served from GPUs in your region behind one endpoint.
Deepseek V4 Pro
DeepSeek
1.6T-parameter flagship MoE (49B active) for maximum reasoning, coding, and agentic performance.
Deepseek V4 Flash
DeepSeek
Cost-efficient V4 variant — 284B parameters (13B active) with the same 1M-token context.
GLM 5.2
Zhipu AI
MIT-licensed 744B MoE with a 1M-token context — tops open-weight coding and design benchmarks.
Kimi K2.6
Moonshot AI
Roughly 1T-parameter MoE with 256K context, tuned for long-horizon tasks and multi-step tool use.
Kimi K2.7 Code
Moonshot AI
Coding-specialized K2 variant that cuts thinking tokens by about 30% versus K2.6.
MiniMax M3
MiniMax
First open-weight model combining frontier coding, 1M context, and native multimodality.
Marketplace
Everything around the model, one click away
Curated tooling for serving, storage, observability, and automation — deployed next to your inference endpoint.