Qwen3 4B vs Qwen3.5 4B

Wins 0 of 2 benchmarks
Wins 1 of 2 benchmarks
Different efforts: non-thinking · thinking
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Qwen3 4B3.3
Qwen3.5 4B9.9
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Qwen3 4B—
Qwen3.5 4B65.4

Specs

SpecQwen3 4BQwen3.5 4B
Input, per 1M$0.40$0.03
Output, per 1M$0.80$0.15
Speed, tok/s—18
Context33K262K
WeightsOpenOpen
ReleasedApr 2025Mar 2026
Memory (4-bit)2 GB2.3 GB
LicenceApache-2.0Apache-2.0
Commercial useAllowedAllowed
See images—✓
Read PDFs?—
Hear audio——
Watch video——
Call tools✓✓
Force a tool call??
Return strict JSONPartialPartial
Search the web??
Run code??
Use a computer??

What breaks when you switch

Qwen3 4B → Qwen3.5 4B
Memory (4-bit): 2 GB → 2.3 GB
Qwen3.5 4B → Qwen3 4B
Context: 262K → 33K
No image input