StreamLake: Models and Prices
Model API platform that serves Kuaishou’s KAT models and open-weight models.
StreamLake served 1.75T tokens of other labs' open-weight models over Nov 2 to Nov 8, 2026 through a large public LLM router, #12 among providers.
Open-weight models
| Model | Family | Share of model | Tokens, 7d | Input $/M | Output $/M | Tokens/s | Uptime 3d | Context |
|---|---|---|---|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017 | DeepSeek | 1.0% | 588B | $0.111 | $0.444 | 62 | 100.0% | 1.02M |
| DeepSeek V4.1 Flash | DeepSeek | 2.7% | 476B | $0.147 | $0.588 | 86 | 100.0% | 1.02M |
| DeepSeek V4 Flash 0423 | DeepSeek | 10.2% | 226B | $0.083 | $0.165 | 24 | 100.0% | 1.02M |
| DeepSeek V4 Pro 0423 | DeepSeek | 26.0% | 130B | $0.894 | $1.79 | 40 | 100.0% | 1.02M |
| DeepSeek V4 Flash 0731 | DeepSeek | 3.8% | 124B | $0.132 | $0.396 | 45 | 100.0% | 1.02M |
| DeepSeek V4 Pro 0813 | DeepSeek | 12.8% | 81B | $0.66 | $1.98 | 58 | 100.0% | 1.02M |
| GLM 5.2 | GLM | 7.2% | 56B | $0.613 | $1.93 | 46 | 100.0% | 1.02M |
| GLM 5.3 Flash | GLM | 0.3% | 44B | $0.088 | $0.295 | 47 | 99.2% | 1.02M |
| MiniMax M3 | MiniMax | 2.8% | 16B | $0.30 | $1.20 | 90.5 | 99.9% | 1M |
| MiniMax M3.1 Flash | MiniMax | 0.2% | 11B | $0.135 | $0.54 | 59 | 99.9% | 1M |
| MiMo-V2.6-Pro | MiMo | - | - | $0.522 | $1.04 | 38 | 93.1% | 1M |
| Qwen3.8 27B | Qwen | - | - | $0.175 | $1.05 | 60 | 100.0% | 1.02M |
| Kimi K2.6 | Kimi | - | - | $0.599 | $2.52 | 41 | 99.2% | 256K |
| Kimi K2.7 Code | Kimi | - | - | $0.713 | $3.00 | 62 | 100.0% | 256K |
| GLM 5 | GLM | - | - | $0.60 | $1.92 | 42 | 99.8% | 198K |
| GLM 5.1 | GLM | - | - | $0.966 | $3.04 | 28 | 100.0% | 200K |
| DeepSeek V3 | DeepSeek | - | - | $0.257 | $1.03 | 46 | 100.0% | 128K |
| Qwen3.5 397B A17B | Qwen | - | - | $0.60 | $3.60 | 51 | 100.0% | 256K |
| R1 0528 | DeepSeek | - | - | $0.571 | $2.29 | 41 | 100.0% | 128K |
| MiniMax M2.5 | MiniMax | - | - | $0.27 | $1.08 | 102.5 | 100.0% | 200K |
List prices per million tokens as of November 8, 2026. Share: of all tokens served on that model; a dash where the volume is not reported.
Company
- Website
- streamlake.com
- Headquarters
- China
- Status page
- status.streamlake.ai
Sources: Artificial Analysis.
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