I posted the updated results here:
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Qwen/Qwen3.6-35B-A3B (and FP8) has landed | 152 | 12768 | April 26, 2026 | |
| Qwen3.5 27B optimisation thread starting at 30+ t/s TP=1 | 18 | 1443 | April 16, 2026 | |
| Qwen3.6-27B-Dflash link | 20 | 943 | April 27, 2026 | |
| Bfloat16 Quality = Speed? | 42 | 1460 | April 26, 2026 | |
| Qwen/Qwen3.5-122B-A10B - Alibaba/Qwen thought about us... :-D | 340 | 15145 | March 24, 2026 | |
| Qwen3.5-122B-A10B on single Spark: up to 51 tok/s (v2.1 — patches + quick-start + benchmark) | 342 | 10848 | April 27, 2026 | |
| HOW-TO: Run Qwen3-Coder-Next on Spark | 92 | 8847 | March 24, 2026 | |
| Introducing PrismaQuant | 110 | 2715 | April 25, 2026 | |
| DFlash LLM for DGX Spark - too good to be true? | 37 | 2186 | April 17, 2026 | |
| What's the best speed we can get with Qwen 3.6 27B without quantizing? | 26 | 4360 | April 24, 2026 |