[whatmodel.sh](../../README.md) / Mellum2 12B-A2.5B Thinking
# Mellum2 12B-A2.5B Thinking
JetBrains · released 2026-05 · 12.15B params (2.5B active) · apache-2.0
**score 54.2** ±1.2 · #300 of 732 · from 18 benchmark results across 2 sources · † no independently run results verified
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Mellum2 12B-A2.5B Thinking among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃●█▄▅▅▄▄▂▁ #267 of 313 24% 5
coding ▁▁▁▃▅▇●▇▅▄▃▁ #239 of 469 36% 5
chat & writing ▁▁▁▃▆▄●█▆▇▇▃ #258 of 402* 4.4% 1
agent tasks · #267 of 313 · 24% of its score
Beats [DeepSeek R1 0528 Qwen3 8B](deepseek-r1-0528-qwen3-8b.md), [Magistral Small 1.0 (2506)](magistral-small-2506.md), [Llama 4 Scout (109B-A17B)](llama-4-scout.md) here, though they rank higher overall. Trails [Granite 4.0 H 1B](granite-4-0-h-1b.md), [Granite 4.0 1B](granite-4-0-1b.md), [SmolLM3 3B](smollm3-3b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
BFCL V4 (December 2025 archive) (… 45.6% ▁▁▁▂▆▆▅▆●▂▄▃ beats 70/114 medium · avg 42% 8.2% at its level · several runs
Terminal-Bench 2.1 (maker) 0.6% ●·▁·▁▁▄▃▆█▅█ beats 0/63 very hard · avg 6.2% 5.1% sharp
SWE-bench Verified (maker) 2.0% ▁●▁▁▁▁▂▄█▆▁▁ beats 1/138 hard · avg 16% 5.0% sharp · noisy
SWE-bench Pro (split unspecified)… 0.0% ●·▁▁▁▁▂▃█▇▂▁ beats 0/58 very hard · avg 7.2% 4.1% sharp · noisy
BFCL v3 (maker) 69.4% ▁▁▁▁▁▃▅█▆●▄▁ beats 148/171 medium · avg 57% 1.7% blunt · at its level
coding · #239 of 469 · 36% of its score
Beats [Qwen3.5 4B](qwen3-5-4b.md), [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md), [Solar Open 100B (102B-A12B)](solar-open-100b.md) here, though they rank higher overall. Trails [QwQ-32B-Preview](qwq-32b-preview.md), [InternLM2.5 20B Chat](internlm2-5-20b-chat.md), [Ministral 3 3B Reasoning 2512](ministral-3-3b-reasoning.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LiveCodeBench v6 (window unspecif… 69.9% ▂▃▄▄▅█●▆▄▇▄▆ beats 37/67 medium · avg 56% 15% at its level · several runs
HumanEval+ (maker) 90.9% ▁·▁▁▂▃▄▄█▅▄● beats 134/145 easy · avg 80% 8.5% sharp
SWE-bench Verified (maker) 2.0% ▁●▁▁▁▁▂▄█▆▁▁ beats 1/138 hard · avg 16% 5.0% sharp · noisy
SWE-bench Pro (split unspecified)… 0.0% ●·▁▁▁▁▂▃█▇▂▁ beats 0/58 very hard · avg 7.2% 4.1% sharp · noisy
MBPP+ (maker) 75.4% ▁··▁▁▃▄█▆●▁▁ beats 101/120 medium · avg 65% 3.5% at its level
chat & writing · #258 of 402 · 4.4% of its score
Beats [Mistral Small 4 (119B-A6.5B)](mistral-small-4.md), [Qwen2.5 Coder 32B Instruct](qwen2-5-coder-32b.md), [Ministral 3 14B Reasoning 2512](ministral-3-14b-reasoning.md) here, though they rank higher overall. Trails [Flan-T5 XXL](flan-t5-xxl.md), [Granite 4.0 H 1B](granite-4-0-h-1b.md), [Granite 4.0 1B](granite-4-0-1b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
IFEval (aggregation unspecified) … 76.5% ▁▂▄▂▂▄●██▅▂▁ beats 51/145 easy · avg 81% 4.4% several runs
other tests (3)
test result vs the others in its run difficulty counts why
GPQA Diamond (maker) 57.6% ▁▁▁▁▂▄▅●███▄ beats 108/322 hard · avg 43% 25% sharp · at its level · several runs
MMLU-Redux (maker) 86.2% ▁▁▃▁▂▃▄●▇▅█▄ beats 24/63 easy · avg 79% 14% several runs
GSM-Plus (maker) 87.0% ▁··▁▁▁·▃▁▄●▁ beats 10/16 easy · avg 84% 5.0% blunt · several runs
source results share of the score
JetBrains/Mellum2.1-12B-A2.5B-Thinking … 11 72% ▇▇▇▇▇▇▇▇▇
JetBrains/Mellum2-12B-A2.5B-Thinking mo… 7 27% ▇▇▇
prior: pull toward the middle 1% ▇
weighted average = 54.2
Licences: JetBrains/Mellum2.1-12B-A2.5B-Thinking model card ([Apache-2.0](https://huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking/blob/92ddae9fc7665e9f801d141d2e5a6b2caf2460c4/README.md)) · JetBrains/Mellum2-12B-A2.5B-Thinking model card ([Apache-2.0](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Thinking/blob/a7311550557e93cc706ab5dd3d879c1a11703ab4/README.md))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run Mellum2 12B-A2.5B Thinking (Q4_K_M, needs ~8.2 GB, from [huggingface.co/bartowski](https://huggingface.co/bartowski/Mellum2-12B-A2.5B-Thinking-GGUF/tree/e10a3d18c7a65d9e06a1ef9c228082fbbc798182))
```sh
llama cli -hf bartowski/Mellum2-12B-A2.5B-Thinking-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf bartowski/Mellum2-12B-A2.5B-Thinking-GGUF:Q4_K_M
```
Other versions: Q8_0 (~12.9 GB) `llama cli -hf bartowski/Mellum2-12B-A2.5B-Thinking-GGUF:Q8_0`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run Mellum2 12B-A2.5B Thinking (Q4_K_M, needs ~8.2 GB, from [huggingface.co/bartowski](https://huggingface.co/bartowski/Mellum2-12B-A2.5B-Thinking-GGUF/tree/e10a3d18c7a65d9e06a1ef9c228082fbbc798182))
```sh
lms get https://huggingface.co/bartowski/Mellum2-12B-A2.5B-Thinking-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q8_0 (~12.9 GB) `lms get https://huggingface.co/bartowski/Mellum2-12B-A2.5B-Thinking-GGUF@Q8_0 -y` · 4bit (~6.8 GB, Mac only) `lms get https://huggingface.co/mlx-community/Mellum2-12B-A2.5B-Thinking-4bit -y` · 8bit (~12.9 GB, Mac only) `lms get https://huggingface.co/mlx-community/Mellum2-12B-A2.5B-Thinking-8bit -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run Mellum2 12B-A2.5B Thinking (Q4_K_M, needs ~8.2 GB, from [huggingface.co/bartowski](https://huggingface.co/bartowski/Mellum2-12B-A2.5B-Thinking-GGUF/tree/e10a3d18c7a65d9e06a1ef9c228082fbbc798182))
```sh
ollama run hf.co/bartowski/Mellum2-12B-A2.5B-Thinking-GGUF:Q4_K_M
```
Other versions: Q8_0 (~12.9 GB) `ollama run hf.co/bartowski/Mellum2-12B-A2.5B-Thinking-GGUF:Q8_0`
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run Mellum2 12B-A2.5B Thinking (4bit, needs ~6.8 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/Mellum2-12B-A2.5B-Thinking-4bit/tree/0821bd05a58a0d5aa5fec32b7738a84c060404dd))
```sh
mlx_lm.chat --model mlx-community/Mellum2-12B-A2.5B-Thinking-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model mlx-community/Mellum2-12B-A2.5B-Thinking-4bit
```
Other versions: 8bit (~12.9 GB, Mac only) `mlx_lm.chat --model mlx-community/Mellum2-12B-A2.5B-Thinking-8bit`
### Other ways to run it
1. Run it as published (BF16, needs ~24.3 GB, NVIDIA only, from [huggingface.co/JetBrains](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Thinking/tree/a7311550557e93cc706ab5dd3d879c1a11703ab4))
```sh
vllm serve JetBrains/Mellum2-12B-A2.5B-Thinking
```
Model repository and documentation: [huggingface.co/JetBrains](https://huggingface.co/JetBrains/Mellum2-12B-A2.5B-Thinking)
[interactive version](https://whatmodel.sh/model/mellum2-12b-a2-5b-thinking) · [JSON](https://whatmodel.sh/api?format=json&all=1)