SiliconFlow Models — Pricing & Benchmarks
13 models available
SiliconFlow hosts 13 AI models in this catalog. The lowest listed input price is Qwen2.5-7B-Instruct at $0.04/1M input tokens. LLM Reference lets you compare these models across all 80 providers without switching tabs.
| Model | Input (per 1M) | Output (per 1M) | Context | |
|---|---|---|---|---|
| Qwen2.5-7B-Instruct | $0.04 | $0.04 | 128k | |
| Qwen2.5-14B-Instruct | $0.08 | $0.08 | 128k | |
| DeepSeek V3 | $0.15 | $0.5 | 64k | |
| Qwen2.5-32B-Instruct | $0.15 | $0.15 | 128k | |
| Qwen2.5-Coder-32B-Instruct | $0.18 | $0.18 | 128k | |
| Mixtral 8x7B | $0.2 | $0.2 | 32k | |
| DeepSeek R1 | $0.25 | $0.8 | 128k | |
| Qwen2.5-72B-Instruct | $0.28 | $0.28 | 128k | |
| Mistral NeMo (2407) | $0.3 | $0.3 | 128k | |
| Grok-2 | $0.5 | $0.5 | 128k | |
| Nex-N2-Pro | $0.5 | $2.50 | 262k | |
| Mixtral 8x22B Instruct v0.1 | $0.65 | $0.65 | 64k | |
| Mistral Large 2 (2407) | $2.00 | $2.00 | 128k |
Where else to run this
Grok-2 on SiliconFlow
Provider setup and pricing
Mistral Large 2 (2407) on SiliconFlow
Provider setup and pricing
Mixtral 8x7B on SiliconFlow
Provider setup and pricing
Mistral Large 2 (2407) on Microsoft Foundry
Alternative host
Mixtral 8x7B on Databricks Foundation Model Serving
Alternative host
DeepSeek R1 on DeepSeek Platform
Alternative host
Pricing Overview
Cheapest$0.04/1M
Most expensive$2.00/1M