PegaFlow External KV Cache for vLLM
PegaFlow external KV cache helps vLLM serving teams preserve and share KV cache across restarts, instances, and RDMA nodes.
Filter posts by category
DeepSeek-V4-Pro on Novita AI: 1M Context, #1 LiveCodeBench Score
DeepSeek-V4-Pro is a 1.6T-parameter open-source MoE model delivering 1 LiveCodeBench score (93.5) and 1M-token context. Available now via Novita AI.
DeepSeek-V4-Flash on Novita AI: Fast Reasoning at Lower Cost
DeepSeek-V4-Flash is available via Novita AI with the July 31, 2026 refresh. Same 284B/13B MoE design, 1M context, and $0.14/$0.28 pricing.
Ling-2.6-1T on Novita AI: Free API, SWE-Bench SOTA, 1T Param Model
Ling-2.6-1T is Ant Group's trillion-scale model built on MLA + Hybrid Linear Attention — not standard MoE. It achieves open-source SOTA on agent benchmarks (SWE-bench, BFCLv4, TAU2
Ling-2.6-flash on Novita AI: 340 Tokens/s, ~7x Token Efficiency
Ling-2.6-flash is a 104B MoE model (7.4B active) delivering 340 tokens/s and 7x better token efficiency than Nemotron-3-Super on agent benchmarks. Available now via OpenRouter with
Kimi K2.6 on Novita AI: API Pricing ($0.95/$4.00), SWE-Bench & Agentic Coding
Kimi K2.6 is now on Novita AI. 1T MoE open-source model, 256K context, 58.6% SWE-Bench Pro — built for long-horizon agentic coding. Try free via OpenAI-compatible API.
GLM-5.1 API on Novita AI: Long-Horizon Agentic Model
GLM-5.1 by Z.ai is now on Novita AI — tops SWE-Bench Pro at 58.4 and runs autonomous coding tasks for 8 hours. Try it with our serverless API, pay per token.
Google Gemma 4 Is Now on Novita AI — 4 Sizes, Audio, Vision Support
Gemma 4 is now available on Novita AI — 4 model sizes, 3 architectures, vision support across the lineup, audio support on E2B and E4B.
Kimi K2.5 Now on Novita AI: Multimodal AI for Vision, Code and Agent
Access Kimi K2.5 on Novita AI - Moonshot AI's flagship multimodal agentic model with 256K context, vision+text, thinking modes & agent swarm.
What New Applications Does Speech 2.6 Enable Compared to Speech 2.5?
Explore the advantages of Speech 2.6 for TTS and voice agents. Discover how it boosts productivity and efficiency in applications.
MiniMax Speech 2.5 Solves Real-Time Multilingual Voice Challenges
Explore MiniMax Speech 2.5, a solution for high-accuracy voice cloning with fast response times and multilingual support.
GLM-4.6V on Novita AI: Vision AI with Native Tool Calling
Access GLM-4.6V API on Novita AI: 106B vision-language model with 128K context, native function calling, and SoTA multimodal document understanding.
MiniMax Speech-2.6 on Novita AI: Next-Gen TTS Model for Voice Synthesis
Try MiniMax Speech-2.6 on Novita AI with faster response, lifelike voice quality, and global language coverage for any voice scenario.
Novita Build Month: Everything you need to build AI applications, now up to 20% off
Get 20% off LLM APIs, GPUs, and Agent Sandbox through Dec 31. Access 200+ AI models, RTX 4090/5090 instances, and autonomous agent infrastructure.