数据状态
同步自 OpenRouter Providers · 101 家托管商 · source=openrouter · 2026-08-01T19:40:54.988Z
SiliconFlow
总部 新加坡 · 机房 US
35 模型38 端点日可用 96.6%
托管模型35
端点38
日可用率均96.6%
最低输入 $/M$0.040
最大上下文1.0M
托管端点 · 38
来自 OpenRouter 各模型 Endpoints API：同一模型可能有多量化 / 多区域端点。
| 模型 | Tag | 量化 | 上下文 | 输入 | 输出 | 5m | 30m | 1d | 延迟 | 吞吐 | 缓存 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| step-3.5-flash stepfun/step-3.5-flash | siliconflow/fp8 | fp8 | 262k | $0.100 | $0.300 | 100.00% | 100.00% | 100.00% | — | — | — |
| hunyuan-a13b-instruct tencent/hunyuan-a13b-instruct | siliconflow/fp8 | fp8 | 131k | $0.140 | $0.570 | — | — | 100.00% | — | — | — |
| minimax-m2.5 minimax/minimax-m2.5 | siliconflow/fp8 | fp8 | 197k | $0.300 | $1.20 | — | 100.00% | 100.00% | — | — | — |
| glm-5.1 z-ai/glm-5.1 | siliconflow/fp8 | fp8 | 205k | $1.19 | $3.74 | 100.00% | 100.00% | 99.96% | — | — | — |
| nex-n2-pro nex-agi/nex-n2-pro | siliconflow | — | 262k | $0.500 | $2.50 | — | 100.00% | 99.95% | — | — | — |
| kimi-k2.7-code moonshotai/kimi-k2.7-code | siliconflow/fp8 | fp8 | 262k | $0.859 | $3.80 | — | — | 99.94% | — | — | — |
| glm-5.2 z-ai/glm-5.2 | siliconflow/fp8 | fp8 | 1.0M | $1.19 | $3.74 | 100.00% | 100.00% | 99.8% | — | — | — |
| deepseek-v4-flash-0731 deepseek/deepseek-v4-flash-0731 | siliconflow/fp8 | fp8 | 1.0M | $0.140 | $0.280 | 100.00% | 99.6% | 99.7% | — | — | — |
| glm-5 z-ai/glm-5 | siliconflow/fp8 | fp8 | 205k | $0.950 | $2.55 | 100.00% | 99.8% | 99.6% | — | — | — |
| deepseek-v3.2-exp deepseek/deepseek-v3.2-exp | siliconflow/fp8 | fp8 | 164k | $0.270 | $0.410 | 100.00% | 100.00% | 99.6% | — | — | — |
| deepseek-v4-flash deepseek/deepseek-v4-flash | siliconflow/fp8 | fp8 | 1.0M | $0.130 | $0.280 | 99.6% | 99.6% | 99.5% | — | — | — |
| glm-4.5-air z-ai/glm-4.5-air | siliconflow/fp8 | fp8 | 131k | $0.140 | $0.860 | 99.0% | 96.0% | 99.5% | — | — | — |
| gemma-4-26b-a4b-it google/gemma-4-26b-a4b-it | siliconflow/fp8 | fp8 | 262k | $0.120 | $0.400 | 99.7% | 99.91% | 99.4% | — | — | — |
| gemma-4-26b-a4b-it google/gemma-4-26b-a4b-it | siliconflow/fp8 | fp8 | 262k | $0.120 | $0.400 | 99.7% | 99.91% | 99.4% | — | — | — |
| deepseek-v4-pro deepseek/deepseek-v4-pro | siliconflow/fp8 | fp8 | 1.0M | $1.50 | $3.14 | 99.4% | 99.6% | 99.3% | — | — | — |
| qwen3-32b qwen/qwen3-32b | siliconflow/fp8 | fp8 | 131k | $0.140 | $0.570 | 100.00% | 99.8% | 99.3% | — | — | — |
| deepseek-v3.2 deepseek/deepseek-v3.2 | siliconflow/fp8 | fp8 | 164k | $0.259 | $0.420 | 99.9% | 99.9% | 99.2% | — | — | — |
| qwen3.5-35b-a3b qwen/qwen3.5-35b-a3b | siliconflow/fp8 | fp8 | 262k | $0.240 | $1.80 | — | 100.00% | 98.8% | — | — | — |
| qwen3-coder-30b-a3b-instruct qwen/qwen3-coder-30b-a3b-instruct | siliconflow/fp8 | fp8 | 262k | $0.070 | $0.280 | 97.9% | 98.2% | 98.3% | — | — | — |
| deepseek-r1-0528 deepseek/deepseek-r1-0528 | siliconflow/fp8 | fp8 | 164k | $0.500 | $2.18 | 100.00% | 100.00% | 97.8% | — | — | — |
| qwen3-vl-30b-a3b-instruct qwen/qwen3-vl-30b-a3b-instruct | siliconflow/fp8 | fp8 | 262k | $0.290 | $1.00 | 98.7% | 98.7% | 97.6% | — | — | — |
| qwen3.5-27b qwen/qwen3.5-27b | siliconflow/fp8 | fp8 | 262k | $0.250 | $2.00 | 98.7% | 97.9% | 96.9% | — | — | — |
| qwen3.6-35b-a3b qwen/qwen3.6-35b-a3b | siliconflow/fp8 | fp8 | 262k | $0.200 | $1.60 | 100.00% | 99.9% | 96.9% | — | — | — |
| qwen3-vl-30b-a3b-thinking qwen/qwen3-vl-30b-a3b-thinking | siliconflow/fp8 | fp8 | 262k | $0.290 | $1.00 | — | — | 96.6% | — | — | — |
| qwen3.6-27b qwen/qwen3.6-27b | siliconflow/fp8 | fp8 | 262k | $0.300 | $3.20 | 100.00% | 98.5% | 96.6% | — | — | — |
| gpt-oss-20b openai/gpt-oss-20b | siliconflow/fp8 | fp8 | 131k | $0.040 | $0.180 | 100.00% | 94.6% | 96.0% | — | — | — |
| gpt-oss-20b openai/gpt-oss-20b | siliconflow/fp8 | fp8 | 131k | $0.040 | $0.180 | 100.00% | 94.6% | 96.0% | — | — | — |
| qwen3.5-9b qwen/qwen3.5-9b | siliconflow/fp8 | fp8 | 262k | $0.100 | $0.150 | 84.7% | 89.1% | 96.0% | — | — | — |
| qwen3-30b-a3b-instruct-2507 qwen/qwen3-30b-a3b-instruct-2507 | siliconflow/fp8 | fp8 | 262k | $0.090 | $0.300 | 100.00% | 98.6% | 95.5% | — | — | — |
| qwen3.5-122b-a10b qwen/qwen3.5-122b-a10b | siliconflow/fp8 | fp8 | 262k | $0.260 | $2.08 | 100.00% | 100.00% | 95.0% | — | — | — |
| deepseek-v3.1-terminus deepseek/deepseek-v3.1-terminus | siliconflow/fp8 | fp8 | 164k | $0.270 | $1.00 | 100.00% | 99.5% | 94.9% | — | — | — |
| kimi-k2.6 moonshotai/kimi-k2.6 | siliconflow/fp8 | fp8 | 262k | $0.770 | $3.40 | — | 95.6% | 94.8% | — | — | — |
| kimi-k2.5 moonshotai/kimi-k2.5 | siliconflow/int4 | int4 | 262k | $0.450 | $2.25 | 100.00% | 86.2% | 94.5% | — | — | — |
| deepseek-chat-v3.1 deepseek/deepseek-chat-v3.1 | siliconflow/fp8 | fp8 | 164k | $0.270 | $1.00 | 100.00% | 99.1% | 91.9% | — | — | — |
| gemma-4-31b-it google/gemma-4-31b-it | siliconflow/fp8 | fp8 | 262k | $0.130 | $0.400 | 77.6% | 94.4% | 90.2% | — | — | — |
| gemma-4-31b-it google/gemma-4-31b-it | siliconflow/fp8 | fp8 | 262k | $0.130 | $0.400 | 77.6% | 94.4% | 90.2% | — | — | — |
| deepseek-chat-v3-0324 deepseek/deepseek-chat-v3-0324 | siliconflow/fp8 | fp8 | 164k | $0.250 | $1.00 | — | — | 89.9% | — | — | — |
| gpt-oss-120b openai/gpt-oss-120b | siliconflow/fp8 | fp8 | 131k | $0.050 | $0.450 | 96.6% | 92.9% | 73.6% | — | — | — |
覆盖模型 · 35
step-3.5-flashhunyuan-a13b-instructminimax-m2.5glm-5.1nex-n2-prokimi-k2.7-codeglm-5.2deepseek-v4-flash-0731glm-5deepseek-v3.2-expdeepseek-v4-flashglm-4.5-airgemma-4-26b-a4b-itdeepseek-v4-proqwen3-32bdeepseek-v3.2qwen3.5-35b-a3bqwen3-coder-30b-a3b-instructdeepseek-r1-0528qwen3-vl-30b-a3b-instructqwen3.5-27bqwen3.6-35b-a3bqwen3-vl-30b-a3b-thinkingqwen3.6-27bgpt-oss-20bqwen3.5-9bqwen3-30b-a3b-instruct-2507qwen3.5-122b-a10bdeepseek-v3.1-terminuskimi-k2.6kimi-k2.5deepseek-chat-v3.1gemma-4-31b-itdeepseek-chat-v3-0324gpt-oss-120b
元数据与端点观测同步自 OpenRouter 公开 API。GrokCode 不做路由转发；自建测速请用 API 实验室。
站内推荐
同系工具：Clavue CLI / imux IDE / clavue-2.1
Clavue · CLI 与 Agent 平台
CLI / GUI Agent 运行时、设备登录、会员额度与 OpenAI 兼容 API（api.clavue.com）。开发脚本、CI 与本地工具一条链路。
打开 Clavue ↗imux · 原生 macOS AI IDE
Clavue 平台的原生 IDE：多 Agent 分屏、Ghostty 级终端、Agent Chat、浏览器自动化与 Supervisor——不是又一个 Electron 壳。
了解 imux ↗Clavue 2.1 · 产品级大模型
旗舰模型 clavue-2.1（及 fast / pro / rev）：适合复杂推理与长程 Agent 循环。Chat、imux、API 共用会员额度。
查看模型 ↗