vLLM x Novita AI: Chord W4A16 INT4 MoE Kernel for Kimi K2.x
Novita AI's Chord is a W4A16 INT4 MoE CUDA kernel for Kimi K2.x, with Humming-compatible indexed and grouped SM90 paths for vLLM serving.
Filter posts by category
Deploy Hermes Agent in Minutes with Novita Sandbox Template
Deploy Hermes Agent on Novita Agent Sandbox using the Hermes template. Get a persistent, self-improving AI agent running in minutes with no server setup.
NovitaClaw: Run OpenClaw in the Cloud with One Command
Deploy OpenClaw as a persistent 24/7 AI agent on Novita Sandbox with one CLI command. No runtime limits, full model control, and multi-channel support.
How StarSling Builds Self-Improving CI With Novita Agent Sandbox
StarSling uses Novita Agent Sandbox to run CI optimization agents in isolated microVMs, test changes reproducibly, and scale AI workflows across customer pipelines.
DeepSeek Harness With Novita AI: Setup Guide, API Key, and Model Choice
Set up DeepSeek Harness with the Novita AI OpenAI-compatible API. Install the Web UI, add a custom provider, and choose models for coding agents and research.
DeepSeek V4 Flash Vision Exp on Novita AI: Text-Matched Reasoning Gets Image Understanding
DeepSeek V4 Flash Vision Exp on Novita AI is an experimental vision-enabled DeepSeek V4 Flash API with 1M context, image input, and $0.44/$1.32 pricing.
Ling-3.0 Family on Novita AI: Full Lineup Comparison and Which Variant to Use
Compare four Ling-3.0 models on Novita AI by fit, active parameters, context, modalities, and current pricing to choose the right serverless variant.
Ling 3.0 Flash VL API Quick Start for Image and Video Inputs
Send image and video-derived inputs to Ling 3.0 Flash VL through Novita AI's OpenAI-compatible chat completions API.
Ling-3.0-Flash-VL on Novita AI: Native Multimodal API and Pricing
Explore Ling-3.0-Flash-VL on Novita AI: native image and video input, 262K context, 32K output, function calling, and free access through September 23, 2026.
Ling 3.0 Flash Sante on Novita AI: Free API, Specs, and Pricing
See Ling 3.0 Flash Sante's serverless API specs, 256K context, function calling, and time-limited free input/output pricing on Novita AI.
Ling 3.0 Flash Sante Quick Start: First Chat API Call
Send your first Ling 3.0 Flash Sante chat-completions request with the correct Novita model ID, endpoint, Python/cURL examples, and limits.
LLM API Services Compared: Direct, Unified, Gateway, or Self-Hosted
Compare four LLM API services and when each fits: direct provider APIs, unified APIs, gateways, or self-hosted open-model endpoints.
GLM-5.3 Flash on Novita AI: Native Multimodal Coding for Long-Horizon Agents
Explore GLM-5.3 Flash on Novita AI: model ID, 1M context, 128K output, multimodal input, pricing, and where it fits for coding agents.
Ling-3.0-flash-Fin on Novita AI: Finance-Tuned MoE API and Pricing
Explore Ling-3.0-flash-Fin on Novita AI: a finance-enhanced 124B MoE with 256K context, 32K max output, and $0 per-million-token pricing.