DOS
FeaturedDOS is a curated set of cost-efficient models selected for DOSClaw agents and first-party products. The selected model list is published from the live catalog so the stable API ID can evolve without breaking integrations.
Curated Architecture & Routing Engine
Multi-Engine Curated PoolUnified version-stable endpoint pre-tuned for DOSClaw agent workflows, tool calling, and high-performance inference across in-house and partner infrastructure.
Selected Models/Curated Model Pool
4 Curated ModelsCurated set of cost-efficient models pre-selected for DOSClaw agents, backed by in-house GPUs and platform grants.
Qwen3.8-27B is a dense reasoning and coding model served by DOS.AI on its own GPU infrastructure in Singapore. It supports multilingual prompts, tool calling, code generation, structured output, and streaming. Current input and output rates are shown from the live catalog.
GPT-6 Luna is OpenAI's most cost-efficient frontier reasoning model, offering high-speed agentic execution, native function calling, vision, and a 1.05M-token context window. Available through the DOS.AI gateway with Azure Foundry integration and unified billing.
Gemini 3.8 Flash is Google's newest Gemini Flash model, featuring hybrid reasoning, 1M token context window, advanced multimodal capabilities, and sub-second execution. Available through the DOS.AI gateway with unified billing and OpenAI compatibility.
DeepSeek V4.1 Flash is the newest DeepSeek release, a Mixture-of-Experts model with a 1M-token context and controllable thinking (reasoning_effort 1-100). It is available through the DOS.AI gateway on Alibaba Cloud Model Studio with unified billing and OpenAI compatibility.
Core Architectural Pillars
Stable API Identifier
Applications code against model: "dos". Your product never breaks when vendor checkpoints iterate, deprecate, or rename model versions.
DOSClaw Agent Pre-Tuning
Pre-tuned for multi-step agent loops, native JSON schema mode, and ReAct function calling.
Singapore Low-Latency Edge
Direct high-speed NVLink interconnects on Singapore GPU nodes (asia-southeast1) delivering sub-250ms TTFT for high-frequency agent loops and interactive tool execution.
Zero-Markup List Pricing
Billed at the official published list price ($0.07 / 1M in, $0.50 / 1M out) - 100% price passthrough with zero hidden fees.
API Integration
Use the DOS API to integrate DOS into your applications. Compatible with standard OpenAI SDKs for easy drop-in migration.
curl https://api.dos.ai/v1/chat/completions \
-H "Authorization: Bearer $DOS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "dos",
"messages": [
{
"role": "user",
"content": "Hello!"
}
]
}'