SiliconFlow

Models served by SiliconFlow

Model	Context	Input /M	Output /M	Cache read /M	Cache hit	Latency	Throughput
DeepSeek R1-0528 FP8 deepseek-ai/DeepSeek-R1-0528:siliconflow	128k	$0.53/M	$2.29/M	N/A	39.9%	1.7s	19 tps
DeepSeek V3 0324 FP8 deepseek-v3-0324:siliconflow	128k	$0.26/M	$1.05/M	N/A	74.3%	2s	23 tps
DeepSeek V3.1 FP8 deepseek-ai/DeepSeek-V3.1:siliconflow	128k	$0.28/M	$1.05/M	N/A	29.4%	1.3s	14 tps
DeepSeek V3.1 Terminus FP8 deepseek-ai/DeepSeek-V3.1-Terminus:siliconflow	128k	$0.28/M	$1.05/M	N/A	62.3%	2.2s	17 tps
DeepSeek V3.1 Terminus (Thinking) FP8 deepseek-ai/DeepSeek-V3.1-Terminus:thinking:siliconflow	128k	$0.28/M	$1.05/M	N/A	73.1%	2.2s	17 tps
DeepSeek V3.2 FP8 deepseek/deepseek-v3.2:siliconflow	163k	$0.27/M	$0.44/M	$0.14/M	3.9%	2.4s	17 tps
DeepSeek V3.2 (Thinking) FP8 deepseek/deepseek-v3.2:thinking:siliconflow	163k	$0.27/M	$0.44/M	$0.14/M	37%	2.4s	17 tps
DeepSeek V4 Flash FP8 deepseek/deepseek-v4-flash:siliconflow	1.0M	$0.15/M	$0.29/M	$0.03/M	N/A	1.8s	31 tps
DeepSeek V4 Flash (Thinking) FP8 deepseek/deepseek-v4-flash:thinking:siliconflow	1.0M	$0.15/M	$0.29/M	$0.03/M	N/A	1.8s	31 tps
DeepSeek V4 Flash 0731 FP8 deepseek/deepseek-v4-flash-0731:siliconflow	1.0M	$0.23/M	$0.69/M	$0.03/M	N/A	1s	39 tps
DeepSeek V4 Flash 0731 (Thinking) FP8 deepseek/deepseek-v4-flash-0731:thinking:siliconflow	1.0M	$0.23/M	$0.69/M	$0.03/M	N/A	1s	39 tps
DeepSeek V4 Flash Vision Exp FP8 deepseek/deepseek-v4-flash-vision-exp:siliconflow	1.0M	$0.46/M	$1.39/M	$0.03/M	N/A	1.6s	78.5 tps
DeepSeek V4 Pro FP8 deepseek/deepseek-v4-pro:siliconflow	1.0M	$1.83/M	$3.65/M	$0.15/M	N/A	1.9s	36 tps
DeepSeek V4 Pro (Thinking) FP8 deepseek/deepseek-v4-pro:thinking:siliconflow	1.0M	$1.83/M	$3.65/M	$0.15/M	N/A	1.9s	36 tps
DeepSeek V4 Pro 0813 FP8 deepseek/deepseek-v4-pro-0813:siliconflow	1.0M	$1.39/M	$4.16/M	$0.05/M	38.9%	1.8s	43 tps
DeepSeek V4 Pro 0813 (Thinking) FP8 deepseek/deepseek-v4-pro-0813:thinking:siliconflow	1.0M	$1.39/M	$4.16/M	$0.05/M	38.9%	1.8s	43 tps
DeepSeek V4.1 Flash FP8 deepseek/deepseek-v4.1-flash:siliconflow	1.0M	$0.31/M	$1.26/M	$0.0063/M	3.4%	1.4s	91 tps
DeepSeek V4.1 Flash (Thinking) FP8 deepseek/deepseek-v4.1-flash:thinking:siliconflow	1.0M	$0.31/M	$1.26/M	$0.0063/M	76.9%	1.4s	91 tps
Gemma 4 26B A4B FP8 google/gemma-4-26b-a4b-it:siliconflow	262k	$0.15/M	$0.42/M	$0.05/M	N/A	2.1s	18 tps
Gemma 4 26B A4B (Thinking) FP8 google/gemma-4-26b-a4b-it:thinking:siliconflow	262k	$0.15/M	$0.42/M	$0.05/M	N/A	2.1s	18 tps
Gemma 4 31B FP8 google/gemma-4-31b-it:siliconflow	262k	$0.79/M	$1.05/M	$0.26/M	N/A	3.4s	9 tps
Gemma 4 31B (Thinking) FP8 google/gemma-4-31b-it:thinking:siliconflow	262k	$0.79/M	$1.05/M	$0.26/M	N/A	3.4s	9 tps
GLM 4.5 Air FP8 z-ai/GLM-4.5-Air:siliconflow	128k	$0.15/M	$0.90/M	N/A	N/A	1.5s	28 tps
GLM 4.5 Air (Thinking) FP8 z-ai/GLM-4.5-Air:thinking:siliconflow	128k	$0.15/M	$0.90/M	N/A	N/A	1.5s	28 tps
GLM 5 FP8 z-ai/glm-5:siliconflow	200k	$1.00/M	$2.68/M	$0.21/M	74.4%	2.6s	26 tps
GLM 5 (Thinking) FP8 z-ai/glm-5:thinking:siliconflow	200k	$1.00/M	$2.68/M	$0.21/M	0%	2.6s	26 tps
GLM 5.1 FP8 z-ai/glm-5.1:siliconflow	200k	$1.25/M	$3.93/M	$0.63/M	0%	2s	27 tps
GLM 5.1 (Thinking) FP8 z-ai/glm-5.1:thinking:siliconflow	200k	$1.25/M	$3.93/M	$0.63/M	65.6%	2s	27 tps
GLM 5.2 FP8 z-ai/glm-5.2:siliconflow	1.0M	$1.25/M	$3.93/M	$0.23/M	54.2%	1.6s	48 tps
GLM 5.2 (Thinking) FP8 z-ai/glm-5.2:thinking:siliconflow	1.0M	$1.25/M	$3.93/M	$0.23/M	54.2%	1.6s	48 tps
GLM 5.3 FP8 z-ai/glm-5.3:siliconflow	1.0M	$1.18/M	$3.70/M	$0.22/M	0%	1.4s	27.5 tps
GLM 5.3 (Thinking) FP8 z-ai/glm-5.3:thinking:siliconflow	1.0M	$1.18/M	$3.70/M	$0.22/M	1.2%	1.4s	27.5 tps
GLM 5.3 Flash FP8 z-ai/glm-5.3-flash:siliconflow	1.0M	$0.16/M	$0.53/M	$0.03/M	85.7%	2s	24 tps
GPT OSS 120B FP8 openai/gpt-oss-120b:siliconflow	128k	$0.16/M	$0.63/M	$0.08/M	N/A	2s	21.5 tps
GPT OSS 20B FP8 openai/gpt-oss-20b:siliconflow	128k	$0.04/M	$0.19/M	N/A	N/A	1.1s	39 tps
Kimi K2.5 INT4 moonshotai/kimi-k2.5:siliconflow	256k	$0.47/M	$2.36/M	$0.07/M	70%	2.3s	21 tps
Kimi K2.5 Thinking INT4 moonshotai/kimi-k2.5:thinking:siliconflow	256k	$0.47/M	$2.36/M	$0.07/M	48.2%	2.3s	21 tps
Kimi K2.7 Code FP8 moonshotai/kimi-k2.7-code:siliconflow	262k	$0.90/M	$3.99/M	$0.19/M	10.7%	2s	25 tps
LongCat 2.0 longcat-2.0:siliconflow	1.0M	$0.79/M	$3.10/M	$0.02/M	54.8%	6.1s	45.8 tps
LongCat 2.0 Thinking longcat-2.0:thinking:siliconflow	1.0M	$0.79/M	$3.10/M	$0.02/M	85.2%	3.7s	51.5 tps
MiniMax M3 FP8 minimax/minimax-m3:siliconflow	512k	$0.63/M	$2.52/M	$0.13/M	N/A	N/A	N/A
MiniMax M3 Thinking FP8 minimax/minimax-m3:thinking:siliconflow	512k	$0.63/M	$2.52/M	$0.13/M	N/A	N/A	N/A
Qwen 3.6 27B FP8 alibaba/qwen3.6-27b:siliconflow	260k	$0.31/M	$3.36/M	N/A	N/A	1.1s	42 tps
Qwen 3.6 27B Thinking FP8 alibaba/qwen3.6-27b:thinking:siliconflow	260k	$0.31/M	$3.36/M	N/A	N/A	1.1s	42 tps
Qwen 3.6 35B A3B FP8 Qwen/Qwen3.6-35B-A3B:siliconflow	262k	$0.25/M	$1.89/M	$0.16/M	N/A	1.7s	55 tps
Qwen 3.6 35B A3B Thinking FP8 Qwen/Qwen3.6-35B-A3B:thinking:siliconflow	262k	$0.25/M	$1.89/M	$0.16/M	N/A	1.7s	55 tps
Qwen3 32B FP8 qwen/qwen3-32b:siliconflow	41k	$0.15/M	$0.60/M	N/A	0%	1s	28 tps
Qwen3.8 Max FP8 qwen3.8-max:siliconflow	1.0M	$2.10/M	$6.30/M	$0.26/M	N/A	1.3s	40 tps
Qwen3.8 Max Thinking FP8 qwen3.8-max:thinking:siliconflow	1.0M	$2.10/M	$6.30/M	$0.26/M	N/A	1.3s	40 tps
Tencent Hy3 tencent/hy3	262k	$0.07/M	$0.26/M	$0.03/M	N/A	N/A	N/A

Prices are per million tokens. Selectable routes include the applicable provider-selection markup; fixed routes show the standard model price. Availability and pricing refresh continuously; each model page shows the live provider comparison.

Using SiliconFlow via the API

For models with provider selection, append :siliconflow to the model ID, set "provider": "siliconflow" in the request body, or send an X-Provider: siliconflow header. Explicit provider selection adds a route-specific markup over that provider's base price; the prices in the table above already include it. Models without provider selection use the displayed model ID without a provider suffix or selection surcharge.

curl https://nano-gpt.com/api/v1/chat/completions \
  -H "Authorization: Bearer $NANOGPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-ai/DeepSeek-R1-0528:siliconflow",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

See the API documentation for provider preferences, price-aware routing, and error behavior.