Model comparison

Qwen3-4B vs Qwen3.5 27B

Qwen3.5 27B is the stronger model overall, scoring 41.9 to 31.9 on the Noometry Index.

Last verified October 12, 2026. 2 shared benchmarks.

Qwen3-4B

Alibaba (Qwen)

31.9

Rank #264 Confirmed

Qwen3.5 27B

Alibaba (Qwen)

41.9

Rank #127 Confirmed

Summary

They share 2 benchmarks with published results for both. Qwen3-4B scores higher in 0 categories and Qwen3.5 27B in 3 categories; 3 gaps are clear of the uncertainty.
The widest gap is in math, where Qwen3.5 27B leads 38.8 to 29.7.
The biggest single-benchmark swing is MathArena Final-Answer Competitions: 38.5% for Qwen3-4B and 56.7% for Qwen3.5 27B.

Side by side

Qwen3-4B and Qwen3.5 27B specifications
	Qwen3-4B	Qwen3.5 27B
Provider	Alibaba (Qwen)	Alibaba (Qwen)
Noometry Index	31.9	41.9
Released	2025-04-29	2026-02-23
Weights	Open	Open
Context window	—	262K
Max output	—	66K
Input $ / M tokens	—	$0.30
Output $ / M tokens	—	$2.40
Results tracked	6	28

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Not comparable

Qwen3-4B: —, Qwen3.5 27B: 38.9 (#168)

Coding benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
LMArena WebDev	—	1358
WeirdML	—	39.5%
LMArena Coding	—	1427
ALE-Bench	—	349.45

Agentic & Tool Use Not comparable

Qwen3-4B: 27.6 (#100), Qwen3.5 27B: —

Agentic & Tool Use benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
Berkeley Function Calling Leaderboard	35.7%	—
Vending-Bench 2	—	201.98

Reasoning Qwen3.5 27B leads

Qwen3-4B: 19.2 (#268), Qwen3.5 27B: 27.5 (#117)

Reasoning benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
NYT Connections (extended)	—	47.9%
Chess Puzzles	4%	—
Thematic Generalization	—	45.5%
LMArena Hard Prompts	—	1414
DTBench	—	82.4%
LMCA	—	34%

Math Qwen3.5 27B leads

Qwen3-4B: 29.7 (#240), Qwen3.5 27B: 38.8 (#127)

Math benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
MathArena Final-Answer Competitions	38.5%	56.7%
OTIS Mock AIME 2024-2025	52.2%	—
LMArena Math	—	1429

Knowledge Qwen3.5 27B leads

Qwen3-4B: 33.0 (#208), Qwen3.5 27B: 38.0 (#150)

Knowledge benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
Vectara Hallucination Rate	5.7%	12.1%
GPQA Diamond	52.3%	—
LMArena Expert	—	1428

Multimodal Not comparable

Qwen3-4B: —, Qwen3.5 27B: 39.4 (#59)

Multimodal benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
LMArena Vision	—	1241

Multilingual Not comparable

Qwen3-4B: —, Qwen3.5 27B: 50.8 (#115)

Multilingual benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
LMArena Non-English	—	1390
LMArena Chinese	—	1478
LMArena French	—	1410
LMArena German	—	1393
LMArena Japanese	—	1345
LMArena Korean	—	1358
LMArena Russian	—	1390
LMArena Spanish	—	1407

Instruction Following Not comparable

Qwen3-4B: —, Qwen3.5 27B: 73.5 (#119)

Instruction Following benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
LMArena Instruction Following	—	1393

Long Context Not comparable

Qwen3-4B: —, Qwen3.5 27B: 43.1 (#106)

Long Context benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
LMArena Longer Query	—	1413

Writing & Preference Not comparable

Qwen3-4B: —, Qwen3.5 27B: 59.3 (#111)

Writing & Preference benchmarks
Benchmark	Qwen3-4B	Qwen3.5 27B
LMArena Text	—	1409
LMArena Creative Writing	—	1362
LMArena Multi-Turn	—	1410

Frequently asked questions

Is Qwen3-4B better than Qwen3.5 27B?