SiliconFlow

inference-hosting

Properties

vendor: SiliconFlow
url: https://siliconflow.com/
whatItIs: Full-stack AI cloud for serverless inference, fine-tuning, and GPU compute on 200+ LLM and multimodal models, with a built-in AI Gateway for unified routing, rate limiting, and cost control.
hosting: managed-only
pricingModel: per-token (DeepSeek-V4-Flash at $0.13/M input; free tier on select models) + reserved GPU options
modelProviders: open-weight (DeepSeek, Qwen, GLM, Llama, multimodal)
stateModel: stateless
toolModel: OpenAI-compatible API + AI Gateway
includesBrowser: false
includesMemory: false
longRunning: per-request + batch jobs
hitlSupport: false
observability: usage dashboard + cost tracking
maturity: ga
launched: 2023
notes: Leading inference provider for Chinese open-weight models (DeepSeek, GLM, Qwen). Bundles serverless inference, one-click fine-tuning, and a gateway layer in one platform. Recognised as top-5 LLM inference provider in 2026 comparisons.

inference-hosting

vendor: SiliconFlow
url: https://siliconflow.com/
whatItIs: Full-stack AI cloud for serverless inference, fine-tuning, and GPU compute on 200+ LLM and multimodal models, with a built-in AI Gateway for unified routing, rate limiting, and cost control.
hosting: managed-only
pricingModel: per-token (DeepSeek-V4-Flash at $0.13/M input; free tier on select models) + reserved GPU options
modelProviders: open-weight (DeepSeek, Qwen, GLM, Llama, multimodal)
stateModel: stateless
toolModel: OpenAI-compatible API + AI Gateway
includesBrowser: false
includesMemory: false
longRunning: per-request + batch jobs
hitlSupport: false
observability: usage dashboard + cost tracking
maturity: ga
launched: 2023
notes: Leading inference provider for Chinese open-weight models (DeepSeek, GLM, Qwen). Bundles serverless inference, one-click fine-tuning, and a gateway layer in one platform. Recognised as top-5 LLM inference provider in 2026 comparisons.