SiliconFlow
Inference providerChinaprices checked Oct 23, 06:00 UTC
SiliconFlow serves 39 models at a verified price. On the 39 it shares with other hosts, it is the cheapest on 6 and charges a median 62% above the cheapest.
Models served3939 with open weights
Cheapest input, per M$0.04gpt-oss-20b
Uptime, last 24 h97.0%average of its 38 OpenRouter endpoints
Sign-upCard requiredno free credits stated
| Model | Precision | Context | Tools | Input | Output | Cache |
|---|---|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017 | fp8 | 1M | ✓ | $0.30 | $1.20 | $0.006 |
| DeepSeek V4.1 Flash | fp8 | 1M | ✓ | $0.30 | $1.20 | $0.006 |
| GLM-5.3-Flash | fp8 | 1M | ✓ | $0.15 | $0.50 | $0.03 |
| DeepSeek V4 Flash Vision Exp | fp8 | 1M | ✓ | $0.44 | $1.32 | $0.028 |
| GLM-5.3 | fp8 | 1M | ✓ | $0.70 | $2.20 | $0.13 |
| DeepSeek V4 Pro 0813 | fp8 | 1M | ✓ | $1.32 | $3.96 | $0.044 |
| Qwen3.8 2.4T A95Bcheapest | fp8 | 1M | ✓ | $2.00 | $6.00 | $0.25 |
| DeepSeek V4 Flash 0731 | fp8 | 1M | ✓ | $0.22 | $0.66 | $0.028 |
| GLM 5.2 | fp8 | 1M | ✓ | $0.70 | $2.20 | $0.13 |
| Kimi K2.7 Code | fp8 | 256K | ✓ | $0.86 | $3.80 | $0.18 |
| DeepSeek V4 Flash 0423 | fp8 | 1M | ✓ | $0.13 | $0.28 | $0.028 |
| DeepSeek V4 Pro 0423 | fp8 | 1M | ✓ | $1.50 | $3.13 | $0.14 |
| Qwen3.6 27B | fp8 | 256K | ✓ | $0.30 | $3.20 | — |
| Kimi K2.6 | fp8 | 256K | ✓ | $0.77 | $3.40 | $0.14 |
| Qwen3.6 35B A3B | fp8 | 256K | — | $0.24 | $1.80 | $0.15 |
| GLM 5.1 | fp8 | 200K | ✓ | $1.19 | $3.74 | $0.60 |
| Gemma 4 26B A4B | fp8 | 256K | — | $0.14 | $0.40 | $0.05 |
| Gemma 4 31B | fp8 | 256K | ✓ | $0.75 | $1.00 | $0.25 |
| Qwen3.5 122B A10Bcheapest | fp8 | 256K | — | $0.26 | $2.08 | — |
| Qwen3.5 9Bcheapest | fp8 | 256K | ✓ | $0.10 | $0.15 | — |
| Qwen3.5-27Bcheapest | fp8 | 256K | ✓ | $0.25 | $2.00 | — |
| Qwen3.5-35B-A3B | fp8 | 256K | — | $0.24 | $1.80 | $0.15 |
| GLM 5 | fp8 | 200K | ✓ | $0.95 | $2.55 | $0.20 |
| MiniMax M2.5 | not read: host terms | 200K | — | $0.30 | $1.20 | $0.03 |
| Kimi K2.5 | int4 | 256K | ✓ | $0.50 | $2.50 | $0.15 |
| DeepSeek V3.2 | fp8 | 160K | ✓ | $0.26 | $0.80 | $0.14 |
| Qwen3 VL 30B A3B Thinkingcheapest | fp8 | 256K | ✓ | $0.29 | $1.00 | — |
| Qwen3 VL 30B A3B Instruct | fp8 | 256K | ✓ | $0.29 | $1.00 | — |
| DeepSeek V3.2 Exp | fp8 | 160K | ✓ | $0.27 | $0.41 | — |
| DeepSeek V3.1 Terminuscheapest | fp8 | 160K | ✓ | $0.27 | $1.00 | — |
| DeepSeek V3.1 | fp8 | 160K | ✓ | $0.27 | $1.00 | — |
| gpt-oss-120b | fp8 | 128K | — | $0.15 | $0.60 | $0.075 |
| gpt-oss-20b | fp8 | 128K | — | $0.04 | $0.18 | — |
| Qwen3 Coder 30B A3B Instruct | fp8 | 256K | ✓ | $0.07 | $0.28 | — |
| GLM 4.5 Air | fp8 | 128K | ✓ | $0.14 | $0.86 | — |
| Qwen3 32B | fp8 | under review | ✓ | $0.14 | $0.57 | — |
| Qwen3 30B A3B | fp8 | 256K | ✓ | $0.09 | $0.30 | — |
| DeepSeek V3 0324 | fp8 | 160K | ✓ | $0.25 | $1.00 | — |
| DeepSeek R1 | fp8 | 160K | ✓ | $0.50 | $2.18 | — |
10 of 39 models · prices per million tokens · precision and context: read at SiliconFlow's own catalogue or docs where its terms allow, otherwise as declared to OpenRouter (hover a cell for its source) · tools as SiliconFlow declares them to OpenRouter
| Condition | What it says | Source |
|---|---|---|
| Sign-up | Card required | docs.siliconflow.com |
| Free credits | None stated | — |
| Headquarters region | China | — |
| Price list | Where its per-token prices are published | siliconflow.com |
| API documentation | Its own docs | docs.siliconflow.com |
| Terms of service | Its terms | docs.siliconflow.com |
| Free credits | $1 in free credits to get started | siliconflow.com |
Links checked Oct 8.