Back to leaderboard

Kimi-48B:48B

OLLAMA GGUF

kimi-linear · 49.1B · unknown

Excellent

Aug 22, 2026 · NVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )

qwen3.5-2b-mlx

LM-STUDIO MLX

qwen3_5 · 2B · 4bit

Good

Mar 6, 2026 · Apple M4 Pro

Global Score
90 vs 65
Hardware Fit
96 vs 88
Quality Score
88 vs 55

Hardware

Kimi-48B:48B qwen3.5-2b-mlx
MachineHP HyperX OMEN Transcend Gaming Laptop 14-ga0xxx5CD612DXMLMac mini
CPUNVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )Apple M4 Pro
Cores2014
RAM55 GB64 GB
GPUNVIDIA RTX Spark N1X (6144-core Blackwell RTX GPU), NVIDIA NPUApple M4 Pro
OSMicrosoft Windows 11 Pro 10.0.28000macOS 15.7.4
Archarm64arm64
Power Modebalancedbalanced

Performance

Kimi-48B:48B qwen3.5-2b-mlx
Tokens/sec62.748.9
First chunk397 ms4208 ms
TTFT397 ms4.2 s
Load time16.8 sN/A
Memory usage25.5 GB1.6 GB
Memory %47%3%

HW Fit Score Breakdown

Kimi-48B:48B

Speed
50/50
TTFT
20/20
Memory
26/30

qwen3.5-2b-mlx

Speed
50/50
TTFT
8/20
Memory
30/30

Quality

Kimi-48B:48B

Reasoning
16/20
Coding
19/20
Instruction
16/20
Structured
15/15
Math
12/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong

qwen3.5-2b-mlx

Reasoning
10/20
Coding
7/20
Instruction
13/20
Structured
14/15
Math
3/15
Multilingual
8/10
Reasoning: Adequate Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Poor Multilingual: Strong

Run yours and compare

$ npm install -g metrillm@latest
$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest