LLM API from Russia: Qwen 3.8, gpt-oss, Gemma, Kimi — billed in rubles

Language models over an OpenAI-compatible API billed in rubles, no VPN and no foreign cards: Qwen 3.8 and Qwen 3.6 on our own GPUs in Russia from 7.14 RUB per 1M tokens, gpt-oss-120b, Gemma 4, Kimi K2.6, GLM, DeepSeek. One base_url and an sk- key. Plus embeddings, rerank, speech, OCR and web search.

Part of NeuralDeep — a self-hosted AI hub: an OpenAI-compatible API to LLMs on our own GPUs, an AI agent and web search under one subscription.

Home · API docs · Models · Pricing · Benchmarks

Machine-readable reference for AI agents: llms.txt · full: llms-full.txt

https://neuraldeep.tech/en/models