Model benchmarks
qwen3.6-14b-a3b-fablevibes local LLM performance
As of August 2026, qwen3.6-14b-a3b-fablevibes runs at up to 95.3 tok/s for local inference (best of 3 community benchmark runs across 2 GPUs).
LM Studio
Model size
14B
Peak speed
95.3 tok/s
Average speed
71.4 tok/s
Min memory
8.7 GB
Max context
5,483 tokens
Best quality
77.6
Benchmark runs
3
GPUs tested
2
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.6-14b-a3b-fablevibes has been benchmarked on, ranked by peak token generation speed. Last updated August 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 6800 XT | LM Studio | — | 95.3 tok/s | 84.4 tok/s | 8.7 GB | 5,483 tokens | 40.7 | 2 |
| NVIDIA GeForce RTX 4070 Ti SUPER | LM Studio | — | 45.2 tok/s | 45.2 tok/s | 14.5 GB | 5,127 tokens | 60.4 | 1 |
Frequently asked questions
- How fast is qwen3.6-14b-a3b-fablevibes for local inference?
- Across 3 community benchmark runs, qwen3.6-14b-a3b-fablevibes reaches up to 95.3 tok/s and averages 71.4 tok/s, with the fastest results on AMD Radeon RX 6800 XT.
- How much memory does qwen3.6-14b-a3b-fablevibes need?
- The leanest observed configuration used about 8.7 GB of memory.
- Which tools have been used to run qwen3.6-14b-a3b-fablevibes?
- Benchmarks were submitted using LM Studio. Results are community-contributed and updated as new runs arrive.