Model benchmarks

llama3.1:8b local LLM performance

As of August 2026, llama3.1:8b runs at up to 107.7 tok/s for local inference (best of 7 community benchmark runs across 1 GPU).

OllamaQ4_K_M
ShareRedditX

Model size

8.0B

Peak speed

107.7 tok/s

Average speed

53.0 tok/s

Min memory

5.8 GB

Max context

65,536 tokens

Avg runtime / run

52s

Avg quality

44.9

Benchmark runs

7

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 7 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

TaskP5 (low)AvgP95 (high)
Overall33.744.954.0
Agent Workflow29.254.575.0
Code Generation15.027.439.0
Role Play & Narrative48.959.067.3
Research & Analysis26.938.549.2

Performance by hardware and tool

Every hardware/tool/quantization combination llama3.1:8b has been benchmarked on, ranked by peak token generation speed. Last updated August 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
NVIDIA GeForce RTX 5080OllamaQ4_K_M107.7 tok/s107.7 tok/s15.4 GB65,536 tokens43.41
CPU onlyOllamaQ4_K_M74.0 tok/s43.9 tok/s5.8 GB8,192 tokens45.16

Frequently asked questions

Is llama3.1:8b good for coding?
In our benchmarks, llama3.1:8b scores 27.4/100 for coding. It runs at about 53.0 tok/s, so if you want more speed, Gemma4:E2B/QAT-MTP@131K is faster (~303.9 tok/s) and still scores well for coding (66.9/100).
Is llama3.1:8b good for agentic (tool-using) tasks?
In our benchmarks, llama3.1:8b scores 54.5/100 for agentic workflows. It runs at about 53.0 tok/s, so if you want more speed, Gemma4:E2B/QAT-MTP@131K is faster (~303.9 tok/s) and still scores well for agentic workflows (71.7/100).
How fast is llama3.1:8b for local inference?
Across 7 community benchmark runs, llama3.1:8b reaches up to 107.7 tok/s and averages 53.0 tok/s, with the fastest results on NVIDIA GeForce RTX 5080.
How much memory does llama3.1:8b need?
The leanest observed configuration used about 5.8 GB of memory (quantizations tested: Q4_K_M).
Which tools have been used to run llama3.1:8b?
Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.