SiliconFlow: Models and Prices
Model-as-a-service token platform; reports trillions of tokens a day and 10,000 enterprise customers.
SiliconFlow served 893B tokens of other labs' open-weight models over Oct 3 to Oct 9, 2026 through a large public LLM router, #17 among providers.
Open-weight models
| Model | Family | Share of model | Tokens, 7d | Input $/M | Output $/M | Tokens/s | Uptime 3d | Context |
|---|---|---|---|---|---|---|---|---|
| GLM 5.3 | GLM | 9.4% | 310B | $1.12 | $3.52 | 48 | 99.7% | 1.05M |
| Hy4 preview | Hunyuan | 1.9% | 152B | $0.834 | $2.50 | 40 | 97.5% | 1.05M |
| GLM 5.3 Flash | GLM | 1.3% | 144B | $0.15 | $0.50 | 28 | 97.2% | 1.05M |
| DeepSeek V3.2 | DeepSeek | 37.8% | 107B | $0.259 | $0.80 | 12 | 95.7% | 164K |
| DeepSeek V4.1 Flash | DeepSeek | 0.2% | 63B | $0.30 | $1.20 | 93 | 99.8% | 1.05M |
| DeepSeek V4 Flash 0423 | DeepSeek | 1.1% | 34B | $0.13 | $0.28 | 37 | 99.9% | 1.05M |
| Qwen3.8 2.4T A95B | Qwen | 69.3% | 31B | $2.00 | $6.00 | 36.5 | 100.0% | 1.05M |
| Kimi K2.5 | Kimi | 93.8% | 14B | $0.50 | $2.50 | 24 | 100.0% | 262K |
| GLM 5.2 | GLM | 0.9% | 13B | $1.19 | $3.74 | 84 | 99.2% | 1.05M |
| DeepSeek V4 Flash 0731 | DeepSeek | 0.2% | 11B | $0.22 | $0.66 | 27 | 99.8% | 1.05M |
| DeepSeek V4 Pro 0423 | DeepSeek | 1.0% | 6B | $1.50 | $3.13 | 36 | 99.9% | 1.05M |
| DeepSeek V4 Flash Vision Exp | DeepSeek | 13.5% | 4B | $0.44 | $1.32 | 138 | 99.9% | 1.05M |
| DeepSeek V3 0324 | DeepSeek | 53.8% | 2B | $0.25 | $1.00 | 23 | 99.9% | 164K |
| DeepSeek V3.2 Exp | DeepSeek | 100.0% | 2B | $0.27 | $0.41 | 14 | 100.0% | 164K |
| DeepSeek V4 Pro 0813 | DeepSeek | - | - | $1.32 | $3.96 | 57 | 100.0% | 1.05M |
| gpt-oss-120b | gpt-oss | - | - | $0.15 | $0.60 | 14 | 77.1% | 131K |
| Gemma 4 31B | Gemma | - | - | $0.75 | $1.00 | 24 | 95.3% | 262K |
| Gemma 4 26B A4B | Gemma | - | - | $0.14 | $0.40 | 41 | 94.0% | 262K |
| gpt-oss-20b | gpt-oss | - | - | $0.04 | $0.18 | 48 | 86.9% | 131K |
| Kimi K2.6 | Kimi | - | - | $0.77 | $3.40 | 26 | 100.0% | 262K |
| Qwen3.6 35B A3B | Qwen | - | - | $0.24 | $1.80 | 77 | 91.3% | 262K |
| DeepSeek V3.1 | DeepSeek | - | - | $0.27 | $1.00 | 14 | 95.8% | 164K |
| GLM 5 | GLM | - | - | $0.95 | $2.55 | 38 | 99.7% | 205K |
| Kimi K2.7 Code | Kimi | - | - | $0.859 | $3.80 | 24 | 100.0% | 262K |
| Qwen3 30B A3B Instruct 2507 | Qwen | - | - | $0.09 | $0.30 | 24 | 99.5% | 262K |
| Qwen3.5-9B | Qwen | - | - | $0.10 | $0.15 | 22 | 98.4% | 262K |
| GLM 5.1 | GLM | - | - | $1.19 | $3.74 | 28 | 99.9% | 205K |
| Qwen3.6 27B | Qwen | - | - | $0.30 | $3.20 | 12 | 92.1% | 262K |
| Qwen3 32B | Qwen | - | - | $0.14 | $0.57 | 30 | 98.3% | 131K |
| Qwen3.5-35B-A3B | Qwen | - | - | $0.24 | $1.80 | 58.5 | 81.9% | 262K |
| GLM 4.5 Air | GLM | - | - | $0.14 | $0.86 | 63.5 | 88.4% | 131K |
| Qwen3.5-27B | Qwen | - | - | $0.25 | $2.00 | 15 | 95.7% | 262K |
| R1 0528 | DeepSeek | - | - | $0.50 | $2.18 | 18 | 99.3% | 164K |
| Qwen3.5-122B-A10B | Qwen | - | - | $0.26 | $2.08 | 40 | 97.5% | 262K |
| Qwen3 VL 30B A3B Instruct | Qwen | - | - | $0.29 | $1.00 | 17 | 84.2% | 262K |
| Qwen3 Coder 30B A3B Instruct | Qwen | - | - | $0.07 | $0.28 | 27 | 99.7% | 262K |
| DeepSeek V3.1 Terminus | DeepSeek | - | - | $0.27 | $1.00 | 20 | 99.0% | 164K |
| Step 3.5 Flash | Step | - | - | $0.10 | $0.30 | 69 | 99.3% | 262K |
| Qwen3 VL 30B A3B Thinking | Qwen | - | - | $0.29 | $1.00 | 43 | 60.0% | 262K |
| Hunyuan A13B Instruct | Hunyuan | - | - | $0.14 | $0.57 | 39 | 100.0% | 131K |
List prices per million tokens as of October 9, 2026. Share: of all tokens served on that model; a dash where the volume is not reported.
Company
- Website
- siliconflow.com
- Headquarters
- Singapore
- Latest financing
- Over 2B yuan ($294M) Series B (June 2026), then Series B+ and C rounds; nearly 2.9B yuan raised in 2026.
Sources: Caixin.
Finance your GPUs
Serving tokens on your own GPUs? Tell us the cluster, the models you serve and the contracts behind them, and we will share it with partner funders that fit.
Prefer email? hello@amcompute.com