Model benchmarks
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit local LLM performance
As of August 2026, NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit runs at up to 110.1 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
30B
Peak speed
110.1 tok/s
Average speed
103.4 tok/s
Min memory
31.7 GB
Max context
65,536 tokens
Avg runtime / run
5m 56s
Avg quality
78.3
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 | 75.7 | 78.3 | 80.8 |
| Agent Workflow | 78.0 | 79.7 | 81.3 |
| Code Generation | 65.7 | 66.9 | 68.1 |
| Role Play & Narrative | 80.3 | 83.8 | 87.3 |
| Research & Analysis | 79.0 | 82.7 | 86.5 |
Performance by hardware and tool
Every hardware/tool/quantization combination NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit 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 |
|---|---|---|---|---|---|---|---|---|
| Apple M5 Max | oMLX | — | 110.1 tok/s | 103.4 tok/s | 31.7 GB | 65,536 tokens | 78.3 | 2 |
Benchmark runs
All 2 NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit good for coding?
- In our benchmarks, NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit scores 66.9/100 for coding. It runs at about 103.4 tok/s, so if you want more speed, Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well for coding (82.2/100).
- Is NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit good for agentic (tool-using) tasks?
- In our benchmarks, NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit scores 79.7/100 for agentic workflows. It runs at about 103.4 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~162.4 tok/s) and still scores well for agentic workflows (87.8/100).
- How fast is NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit for local inference?
- Across 2 community benchmark runs, NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit reaches up to 110.1 tok/s and averages 103.4 tok/s, with the fastest results on Apple M5 Max.
- How much memory does NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit need?
- The leanest observed configuration used about 31.7 GB of memory.
- Which tools have been used to run NVIDIA-Nemotron-3.5-Lightning-30B-A3B-8bit?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.