Is unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF good for agentic (tool-using) tasks?

In our benchmarks, unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF scores 55.3/100 for agentic workflows. It runs at about 102.5 tok/s, so if you want more speed, qwen3.8-flash-next-q2_0 is faster (~127.9 tok/s) and still scores well for agentic workflows (85.6/100).

How fast is unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF for local inference?

Across 5 community benchmark runs, unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF reaches up to 139.0 tok/s and averages 102.5 tok/s, with the fastest results on NVIDIA GeForce RTX 3070 Ti.

How much memory does unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF need?

The leanest observed configuration used about 3.9 GB of memory (quantizations tested: Q4_K_M, UD-Q6_K_XL).

Which tools have been used to run unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF?

Benchmarks were submitted using Unsloth Studio, llama.cpp. Results are community-contributed and updated as new runs arrive.

Model benchmarks

unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF local LLM performance

As of October 2026, unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF runs at up to 139.0 tok/s for local inference (best of 5 community benchmark runs across 2 GPUs).

Unsloth Studiollama.cppQ4_K_MUD-Q6_K_XL

ShareReddit X

Model size

Peak speed

139.0 tok/s

Average speed

102.5 tok/s

Avg PP

1270.6 tok/s

Min memory

3.9 GB

Max context

524,288 tokens

Avg output / run

8,610 tokens

Avg runtime / run

1m 30s

Avg quality

44.3

Benchmark runs

GPUs tested

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 5 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	39.4	44.3	51.0
Agent Workflow	29.4	55.3	72.2
Code Generation	7.1	12.7	19.0
Role Play & Narrative	46.1	53.9	66.0
Research & Analysis	43.9	55.3	66.5

Performance by hardware and tool

Every hardware/tool/quantization combination unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF has been benchmarked on, ranked by peak token generation speed. Last updated October 2026.

Hardware	Tool	Quant	Best tok/s	Avg tok/s	Memory	Context	Quality	Runs
NVIDIA GeForce RTX 3070 Ti	Unsloth Studio	Q4_K_M	139.0 tok/s	139.0 tok/s	3.9 GB	8,192 tokens	49.7	1
NVIDIA GeForce RTX 4070	Unsloth Studio	UD-Q6_K_XL	95.3 tok/s	95.3 tok/s	7.3 GB	65,536 tokens	39.4	1
NVIDIA GeForce RTX 4070	llama.cpp	—	95.0 tok/s	92.8 tok/s	n/a	524,288 tokens	44.1	3

Benchmark runs

All 5 unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

Is unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF good for agentic (tool-using) tasks?: In our benchmarks, unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF scores 55.3/100 for agentic workflows. It runs at about 102.5 tok/s, so if you want more speed, qwen3.8-flash-next-q2_0 is faster (~127.9 tok/s) and still scores well for agentic workflows (85.6/100).
How fast is unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF for local inference?: Across 5 community benchmark runs, unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF reaches up to 139.0 tok/s and averages 102.5 tok/s, with the fastest results on NVIDIA GeForce RTX 3070 Ti.
How much memory does unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF need?: The leanest observed configuration used about 3.9 GB of memory (quantizations tested: Q4_K_M, UD-Q6_K_XL).
Which tools have been used to run unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF?: Benchmarks were submitted using Unsloth Studio, llama.cpp. Results are community-contributed and updated as new runs arrive.