Model benchmarks
llama3.1:latest local LLM performance
As of October 2026, llama3.1:latest runs at up to 77.3 tok/s for local inference (best of 2 community benchmark runs across 2 GPUs).
Model size
8.0B
Peak speed
77.3 tok/s
Average speed
53.2 tok/s
Avg PP
2215.0 tok/s
Min memory
4.4 GB
Max context
8,192 tokens
Avg output / run
3,759 tokens
Avg runtime / run
1m 37s
Avg quality
45.9
Benchmark runs
2
GPUs tested
2
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 | 44.5 | 45.9 | 47.3 |
| Agent Workflow | 35.0 | 55.9 | 76.9 |
| Code Generation | 22.2 | 32.8 | 43.3 |
| Role Play & Narrative | 54.5 | 55.1 | 55.6 |
| Research & Analysis | 35.8 | 39.9 | 44.0 |
Performance by hardware and tool
Every hardware/tool/quantization combination llama3.1:latest 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 3060 Ti | Ollama | Q4_K_M | 77.3 tok/s | 77.3 tok/s | 5.4 GB | 8,192 tokens | 47.5 | 1 |
| NVIDIA GeForce RTX 3070 Ti | Ollama | Q4_K_M | 29.1 tok/s | 29.1 tok/s | 4.4 GB | 8,192 tokens | 44.3 | 1 |
Benchmark runs
All 2 llama3.1:latest runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is llama3.1:latest good for coding?
- In our benchmarks, llama3.1:latest scores 32.8/100 for coding. It runs at about 53.2 tok/s, so if you want more speed, nvidia/Qwen3.8-27B-NVFP4 is faster (~63.3 tok/s) and still scores well for coding (82.7/100).
- Is llama3.1:latest good for agentic (tool-using) tasks?
- In our benchmarks, llama3.1:latest scores 55.9/100 for agentic workflows. It runs at about 53.2 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).
- How fast is llama3.1:latest for local inference?
- Across 2 community benchmark runs, llama3.1:latest reaches up to 77.3 tok/s and averages 53.2 tok/s, with the fastest results on NVIDIA GeForce RTX 3060 Ti.
- How much memory does llama3.1:latest need?
- The leanest observed configuration used about 4.4 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run llama3.1:latest?
- Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.