Neura Intelligence Index

AI Model Benchmarks

Qwen3.5-122B-A10B vs Step3-VL-10B — Benchmark Comparison | Neura Market

Back to Leaderboard

Qwen3.5-122B-A10B vs Step3-VL-10B

Q

Qwen3.5-122B-A10B

Alibaba Cloud / Qwen Team 🇨🇳

69.9

#71—

S

StepFun 🇨🇳

70.4

#68—

Category Comparison

Benchmark-by-Benchmark

reasoning

Benchmark	Qwen3.5-122B-A10B	Step3-VL-10B
GPQA Diamond	86.6	—
Humanity's Last Exam	47.5	—

coding

Benchmark	Qwen3.5-122B-A10B	Step3-VL-10B
SWE-bench Verified	72	—

multimodal

Benchmark	Qwen3.5-122B-A10B	Step3-VL-10B
MMMU	83.9	78.1
MMMU-Pro	76.9	—
CharXiv Reasoning	77.2	—
ScreenSpot Pro	70.4	—

general

Benchmark	Qwen3.5-122B-A10B	Step3-VL-10B
MMMLU	86.7	—

agent

Benchmark	Qwen3.5-122B-A10B	Step3-VL-10B
BrowseComp	63.8	—

math

Benchmark	Qwen3.5-122B-A10B	Step3-VL-10B
AIME 2025	—	87.7

Category Winners

coding

Q

Qwen3.5-122B-A10B

65.6

math

S

Step3-VL-10B

65.9

reasoning

Q

Qwen3.5-122B-A10B

83.0

general

Q

Qwen3.5-122B-A10B

63.7

multimodal

S

Step3-VL-10B

77.2

agent

Q

Qwen3.5-122B-A10B

47.5