Model benchmarks
qwen3.8-flash-next-unsloth-ud-q5_k_xl local LLM performance
As of November 2026, qwen3.8-flash-next-unsloth-ud-q5_k_xl runs at up to 54.1 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
54.1 tok/s
Average speed
51.7 tok/s
Avg PP
213.8 tok/s
Min memory
115.8 GB
Max context
65,536 tokens
Avg output / run
36,682 tokens
Avg runtime / run
11m 48s
Avg quality
81.5
Benchmark runs
2
GPUs tested
1
Quality by task
Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 2 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.
| Task | P5 (low) | Avg | P95 (high) |
|---|---|---|---|
| Overall | 81.5 | 81.5 | 81.5 |
| Agent Workflow | 81.7 | 83.9 | 86.1 |
| Code Generation | 64.3 | 69.0 | 73.7 |
| Role Play & Narrative | 85.8 | 89.0 | 92.1 |
| Research & Analysis | 80.3 | 84.1 | 87.9 |
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.8-flash-next-unsloth-ud-q5_k_xl has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7900 XTX | llama.cpp | Q5_K | 54.1 tok/s | 51.7 tok/s | 115.8 GB | 65,536 tokens | 81.5 | 2 |
Benchmark runs
All 2 qwen3.8-flash-next-unsloth-ud-q5_k_xl runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is qwen3.8-flash-next-unsloth-ud-q5_k_xl good for coding?
- In our benchmarks, qwen3.8-flash-next-unsloth-ud-q5_k_xl scores 69.0/100 for coding. It runs at about 51.7 tok/s, so if you want more speed, local-qwen is faster (~60.2 tok/s) and still scores well for coding (84.8/100).
- Is qwen3.8-flash-next-unsloth-ud-q5_k_xl good for agentic (tool-using) tasks?
- In our benchmarks, qwen3.8-flash-next-unsloth-ud-q5_k_xl scores 83.9/100 for agentic workflows. It runs at about 51.7 tok/s, so if you want more speed, granite-4.2-8b-Q4_K_L is faster (~71.6 tok/s) and still scores well for agentic workflows (92.9/100).
- How fast is qwen3.8-flash-next-unsloth-ud-q5_k_xl for local inference?
- Across 2 community benchmark runs, qwen3.8-flash-next-unsloth-ud-q5_k_xl reaches up to 54.1 tok/s and averages 51.7 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
- How much memory does qwen3.8-flash-next-unsloth-ud-q5_k_xl need?
- The leanest observed configuration used about 115.8 GB of memory (quantizations tested: Q5_K).
- Which tools have been used to run qwen3.8-flash-next-unsloth-ud-q5_k_xl?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.