How fast is hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M for local inference?

Across 3 community benchmark runs, hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M reaches up to 143.9 tok/s and averages 142.7 tok/s, with the fastest results on NVIDIA RTX 2000 Ada Generation Laptop GPU.

How much memory does hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M need?

The leanest observed configuration used about 5.4 GB of memory (quantizations tested: Q4_K_M).

Which tools have been used to run hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M?

Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.

Model benchmarks

hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M local LLM performance

As of November 2026, hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M runs at up to 143.9 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).

OllamaQ4_K_M

ShareReddit X

Model size

6.94B

Peak speed

143.9 tok/s

Average speed

142.7 tok/s

Avg PP

3371.2 tok/s

Min memory

5.4 GB

Max context

65,536 tokens

Avg output / run

5,153 tokens

Avg runtime / run

41s

Avg quality

42.4

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 3 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	33.8	42.4	53.6
Agent Workflow	25.7	47.3	72.3
Code Generation	26.8	33.2	41.3
Role Play & Narrative	35.7	45.9	56.5
Research & Analysis	34.7	43.0	54.9

Performance by hardware and tool

Every hardware/tool/quantization combination hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M 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
NVIDIA RTX 2000 Ada Generation Laptop GPU	Ollama	Q4_K_M	143.9 tok/s	142.7 tok/s	5.4 GB	65,536 tokens	42.4	3

Benchmark runs

All 3 hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

How fast is hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M for local inference?: Across 3 community benchmark runs, hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M reaches up to 143.9 tok/s and averages 142.7 tok/s, with the fastest results on NVIDIA RTX 2000 Ada Generation Laptop GPU.
How much memory does hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M need?: The leanest observed configuration used about 5.4 GB of memory (quantizations tested: Q4_K_M).
Which tools have been used to run hf.co/ibm-granite/granite-4.0-h-tiny-GGUF:Q4_K_M?: Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.