How fast is bonsai-2-27b for local inference?

Across 2 community benchmark runs, bonsai-2-27b reaches up to 39.2 tok/s and averages 38.7 tok/s, with the fastest results on NVIDIA GeForce RTX 2070 SUPER.

How much memory does bonsai-2-27b need?

The leanest observed configuration used about 6.9 GB of memory.

Which tools have been used to run bonsai-2-27b?

Benchmarks were submitted using llama-server, llama-server (Bonsai 2 27B PTQ1_0). Results are community-contributed and updated as new runs arrive.

Model benchmarks

bonsai-2-27b local LLM performance

As of October 2026, bonsai-2-27b runs at up to 39.2 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).

llama-serverllama-server (Bonsai 2 27B PTQ1_0)

ShareReddit X

Model size

27B

Peak speed

39.2 tok/s

Average speed

38.7 tok/s

Avg PP

271.8 tok/s

Min memory

6.9 GB

Max context

8,192 tokens

Avg output / run

16,820 tokens

Avg runtime / run

7m 33s

Avg quality

n/a

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. 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	—	n/a	—
Agent Workflow	—	n/a	—
Code Generation	—	n/a	—
Role Play & Narrative	—	n/a	—
Research & Analysis	—	n/a	—

Performance by hardware and tool

Every hardware/tool/quantization combination bonsai-2-27b 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 2070 SUPER	llama-server	—	39.2 tok/s	39.2 tok/s	6.9 GB	8,192 tokens	n/a	1
NVIDIA GeForce RTX 2070 SUPER	llama-server (Bonsai 2 27B PTQ1_0)	—	38.3 tok/s	38.3 tok/s	6.9 GB	8,192 tokens	n/a	1

Benchmark runs

All 2 bonsai-2-27b runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

How fast is bonsai-2-27b for local inference?: Across 2 community benchmark runs, bonsai-2-27b reaches up to 39.2 tok/s and averages 38.7 tok/s, with the fastest results on NVIDIA GeForce RTX 2070 SUPER.
How much memory does bonsai-2-27b need?: The leanest observed configuration used about 6.9 GB of memory.
Which tools have been used to run bonsai-2-27b?: Benchmarks were submitted using llama-server, llama-server (Bonsai 2 27B PTQ1_0). Results are community-contributed and updated as new runs arrive.