Is JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF good for agentic (tool-using) tasks?

In our benchmarks, JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF scores 61.7/100 for agentic workflows. It runs at about 114.9 tok/s, so if you want more speed, Nex-N2.5-mini-MLX-4bit is faster (~137.0 tok/s) and still scores well for agentic workflows (81.2/100).

How fast is JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF for local inference?

Across 5 community benchmark runs, JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF reaches up to 173.3 tok/s and averages 114.9 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.

How much memory does JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF need?

The leanest observed configuration used about 9.9 GB of memory (quantizations tested: Q4_K_M).

Which tools have been used to run JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF?

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

Model benchmarks

JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF local LLM performance

As of October 2026, JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF runs at up to 173.3 tok/s for local inference (best of 5 community benchmark runs across 1 GPU).

Unsloth StudioQ4_K_M

ShareReddit X

Model size

12B

Peak speed

173.3 tok/s

Average speed

114.9 tok/s

Avg PP

2104.0 tok/s

Min memory

9.9 GB

Max context

65,536 tokens

Avg output / run

23,480 tokens

Avg runtime / run

6m 4s

Avg quality

65.2

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 5 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	62.5	65.2	68.4
Agent Workflow	44.7	61.7	76.9
Code Generation	44.4	54.6	66.8
Role Play & Narrative	70.8	75.4	80.8
Research & Analysis	66.2	69.1	72.2

Performance by hardware and tool

Every hardware/tool/quantization combination JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF 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 4070	Unsloth Studio	Q4_K_M	173.3 tok/s	114.9 tok/s	9.9 GB	65,536 tokens	65.2	5

Benchmark runs

All 5 JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

Is JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF good for agentic (tool-using) tasks?: In our benchmarks, JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF scores 61.7/100 for agentic workflows. It runs at about 114.9 tok/s, so if you want more speed, Nex-N2.5-mini-MLX-4bit is faster (~137.0 tok/s) and still scores well for agentic workflows (81.2/100).
How fast is JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF for local inference?: Across 5 community benchmark runs, JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF reaches up to 173.3 tok/s and averages 114.9 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.
How much memory does JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF need?: The leanest observed configuration used about 9.9 GB of memory (quantizations tested: Q4_K_M).
Which tools have been used to run JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF?: Benchmarks were submitted using Unsloth Studio. Results are community-contributed and updated as new runs arrive.