Neura Intelligence Index

AI Model Benchmarks

IBM Granite 4.2 8B vs Qwen3 VL 30B A3B Instruct — Benchmark Comparison | Neura Market

Back to Leaderboard

IBM Granite 4.2 8B vs Qwen3 VL 30B A3B Instruct

I

IBM Granite 4.2 8B

IBM 🇺🇸

37.4

#233—

Q

Qwen3 VL 30B A3B Instruct

Alibaba Cloud / Qwen Team 🇨🇳

37.6

#230—

Category Comparison

Benchmark-by-Benchmark

reasoning

Benchmark	IBM Granite 4.2 8B	Qwen3 VL 30B A3B Instruct
GPQA Diamond	64.1	70.4

math

Benchmark	IBM Granite 4.2 8B	Qwen3 VL 30B A3B Instruct
AIME 2025	86.7	69.3

coding

Benchmark	IBM Granite 4.2 8B	Qwen3 VL 30B A3B Instruct
SWE-bench Verified	47.7	—
SWE-Bench Pro	19.1	—
SciCode	36.1	—

multimodal

Benchmark	IBM Granite 4.2 8B	Qwen3 VL 30B A3B Instruct
MMMU-Pro	—	60.4
CharXiv Reasoning	—	48.9
ScreenSpot Pro	—	60.5

general

Benchmark	IBM Granite 4.2 8B	Qwen3 VL 30B A3B Instruct
SimpleQA	—	27

agent

Benchmark	IBM Granite 4.2 8B	Qwen3 VL 30B A3B Instruct
OSWorld	—	30.3

Category Winners

coding

I

IBM Granite 4.2 8B

11.6

math

I

IBM Granite 4.2 8B

64.2

reasoning

Q

Qwen3 VL 30B A3B Instruct

55.5

general

Q

Qwen3 VL 30B A3B Instruct

33.4

multimodal

Q

Qwen3 VL 30B A3B Instruct

24.0

agent

Q

Qwen3 VL 30B A3B Instruct

23.4