Is Nex-N2.5-mini-mixed-Q2Q3-vision-trial good for agentic (tool-using) tasks?

In our benchmarks, Nex-N2.5-mini-mixed-Q2Q3-vision-trial scores 81.6/100 for agentic workflows. It runs at about 97.3 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-MLX-oQ4e-MTP is faster (~123.7 tok/s) and still scores well for agentic workflows (84.4/100).

How fast is Nex-N2.5-mini-mixed-Q2Q3-vision-trial for local inference?

Across 4 community benchmark runs, Nex-N2.5-mini-mixed-Q2Q3-vision-trial reaches up to 109.5 tok/s and averages 97.3 tok/s, with the fastest results on NVIDIA GeForce RTX 5060.

How much memory does Nex-N2.5-mini-mixed-Q2Q3-vision-trial need?

The leanest observed configuration used about 16.5 GB of memory.

Which tools have been used to run Nex-N2.5-mini-mixed-Q2Q3-vision-trial?

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

Model benchmarks

Nex-N2.5-mini-mixed-Q2Q3-vision-trial local LLM performance

As of September 2026, Nex-N2.5-mini-mixed-Q2Q3-vision-trial runs at up to 109.5 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).

llama.cpp

ShareReddit X

Model size

Unknown

Peak speed

109.5 tok/s

Average speed

97.3 tok/s

Avg PP

818.5 tok/s

Min memory

16.5 GB

Max context

131,072 tokens

Avg output / run

48,928 tokens

Avg runtime / run

Avg quality

83.4

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 4 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	81.5	83.4	85.8
Agent Workflow	78.9	81.6	83.1
Code Generation	73.9	77.5	80.7
Role Play & Narrative	86.4	88.7	90.9
Research & Analysis	81.5	85.9	89.3

Performance by hardware and tool

Every hardware/tool/quantization combination Nex-N2.5-mini-mixed-Q2Q3-vision-trial has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.

Hardware	Tool	Quant	Best tok/s	Avg tok/s	Memory	Context	Quality	Runs
NVIDIA GeForce RTX 5060	llama.cpp	—	109.5 tok/s	97.3 tok/s	16.5 GB	131,072 tokens	83.4	4

Benchmark runs

All 4 Nex-N2.5-mini-mixed-Q2Q3-vision-trial runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

Is Nex-N2.5-mini-mixed-Q2Q3-vision-trial good for agentic (tool-using) tasks?: In our benchmarks, Nex-N2.5-mini-mixed-Q2Q3-vision-trial scores 81.6/100 for agentic workflows. It runs at about 97.3 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-MLX-oQ4e-MTP is faster (~123.7 tok/s) and still scores well for agentic workflows (84.4/100).
How fast is Nex-N2.5-mini-mixed-Q2Q3-vision-trial for local inference?: Across 4 community benchmark runs, Nex-N2.5-mini-mixed-Q2Q3-vision-trial reaches up to 109.5 tok/s and averages 97.3 tok/s, with the fastest results on NVIDIA GeForce RTX 5060.
How much memory does Nex-N2.5-mini-mixed-Q2Q3-vision-trial need?: The leanest observed configuration used about 16.5 GB of memory.
Which tools have been used to run Nex-N2.5-mini-mixed-Q2Q3-vision-trial?: Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.