Home
Street stacks
since Sep 20
DGX Spark
GLM 5.3 FlashDefaultGLM on both
DeepSeek v4.1 Flash
Qwen3.8-Flash-Next45
MiMo-V2.6-Flash
DGX Spark
- GLM 5.3 FlashHer deskAll three
- GLM 5.3 Flash · Qwen3.8-Flash-NextGLM orch on 2×, Qwen worker on 1×
- DeepSeek v4.1 FlashHomogeneous
- GLM-5.2NVFP4
DGX Spark
- GLM 5.3 Flash · Qwen3.8-Flash-NextGLM orch 2× + Qwen worker 2×
- GLM 5.3 FlashHomogeneous
- DeepSeek v4.1 FlashHomogeneous
DGX Spark
- GLM 5.3 Flash · Qwen3.8-Flash-Next · DeepSeek v4.1 FlashHeard mostGLM orch + Qwen worker + DeepSeek scanner
Top 10 models
Slow <30Fair 30+Fast 80+Turbo 160+
Nations
LabsPayback ranked
Full tableUS price. $1 per million. 20% duty. Left is faster.
| 1 | Qwen3.5 2BRTX 5090 32GB | 2.3 yr | 44% |
| 2 | Qwen3.6-35B-A3BM4 Pro | 4.2 yr | 24% |
| 3 | Qwen3.5-4BM4 Pro | 4.3 yr | 23% |
| 4 | Gemma-4 26B-A4B-itRTX 5090 32GB | 5.1 yr | 20% |
| 5 | Qwen3.8-Flash-NextM5 Max | 7.0 yr | 14% |
| 6 | Qwen3.8-27BRTX 5060 Ti 16GB | 7.0 yr | 14% |
| 7 | Qwen3.5-9BM4 Pro | 7.4 yr | 13% |
| 8 | Qwen3-8BM4 Pro | 7.6 yr | 13% |
Liberated
- NEW351Qwen3.5 2BQ4_K_M
- NEW173
Gemma-4 26B-A4B-itQ4_K_M - NEW126.5Qwen3.8-Flash-Next
- ▼394.4Qwen3.6-35B-A3BNVFP4
- ▼385.5Qwen3.6-35B-A3B4bit
- NEW82.8Qwen3.5-4B4bit
- NEW80.9Qwen3.8-Flash-NextNVFP4
- NEW79.5Qwen3.8-Flash-NextEXL3
- NEW72Qwen3.8-27BNVFP4
- NEW65
AliceAI-Foundation-80B-A3BQ4_K_M - NEW50Empero Qwen3.8-35B-A3BQ4_K_M
- NEW49.3Qwen3.5-9B4bit
- ▼1048.3Qwen3-8B4bit
- NEW47MiMo-V2.6-Distill-Qwen-9BQ5_K_M
- NEW45Qwen3.8-Flash-NextFP8
- NEW40.7Qwen3.8-27BUD-Q4_K_XL
- NEW40
Ornith - NEW23Qwen3.8-27BIQ3_XXS
- NEW19Qwen3.8-27BUD-Q2_K_XL
- NEW18.6GLM 5.3 FlashEXL3
- NEW14
Nex-N2.5-miniQ4_K_M - NEW10Qwen3.8-27B TurboFCFusionIQ2_M
