Phi 4 Multimodal Instruct vs Qwen3.5 4B

Microsoft · released Feb 26, 2025

Wins 0 of 3 areas

—

Alibaba · released Mar 2, 2026

Wins 3 of 3 areas

Coding · Reasoning · Images and charts

Qwen3.5 4B is the stronger all-rounder.

Scores updated Sep 17, 2026 · 14 tests both models report · How we compare

Where each one wins

Tests won in each of the three areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.

Images and chartsUnderstanding pictures, charts and video24Qwen3.5 4B4 of 6 tests
ReasoningHard problems that need careful thinking02Qwen3.5 4B2 of 2 tests
CodingWriting and fixing software01Qwen3.5 4B1 of 1 test

Coding rests on a single test.

What it costs

Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.

Phi 4 Multimodal InstructNo current price is tracked.

Qwen3.5 4B$0.03 to read · $0.15 to write$0.18Price from Artificial Analysis

The biggest differences

The tests each model wins by the widest margin, up to three each. Scores are out of 100.

Where Qwen3.5 4B pulls ahead

Harder college exam questions with imagesMMMU-Pro+50.9points aheadQwen3.5 4B65.4Phi 4 Multimodal Instruct14.5
Graduate-level biology, physics and chemistry questionsGPQA Diamond+45.6points aheadQwen3.5 4B77.1Phi 4 Multimodal Instruct31.5
College exam questions with charts, maps and diagramsMMMU+18.3points aheadQwen3.5 4B73.4Phi 4 Multimodal Instruct55.1

Where Phi 4 Multimodal Instruct pulls ahead

Visual perception tasks people solve at a glanceBLINK+2.6points aheadPhi 4 Multimodal Instruct61.3Qwen3.5 4B58.7
Questions that truly need the image to answerMMStar+1.9points aheadPhi 4 Multimodal Instruct61.2Qwen3.5 4B59.3

Every test, side by side

All 14 tests both models report. The winning score is in its model's colour; marks a score checked independently.

CodingQwen3.5 4B

SciCodeQwen3.5 4B by 5.11116.1+5.1

Full coding ranking

ReasoningQwen3.5 4B

GPQA DiamondQwen3.5 4B by 45.631.577.1+45.6
Humanity's Last ExamQwen3.5 4B by 4.959.9+4.9

Full reasoning ranking

Images and chartsQwen3.5 4B

MMMU-ProQwen3.5 4B by 50.914.565.4+50.9
MMMUQwen3.5 4B by 18.355.173.4+18.3
BLINKPhi 4 Multimodal Instruct by 2.661.358.7+2.6
OCRBenchQwen3.5 4B by 2.584.486.9+2.5
MMStarPhi 4 Multimodal Instruct by 1.961.259.3+1.9
MathVistaQwen3.5 4B by 1.262.463.6+1.2

Full images and charts ranking

Other results5 tests, not counted

Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.

TextVQA-valQwen3.5 4B by 41.339.981.2+41.3
MMBenchPhi 4 Multimodal Instruct by 8.386.778.4+8.3
AA IntelligenceQwen3.5 4B by 7.35.813.1+7.3
DocVQA-valQwen3.5 4B by 292.894.8+2
POPEtie85.686tie

Questions people ask

Which is better, Phi 4 Multimodal Instruct or Qwen3.5 4B?

Qwen3.5 4B wins all three areas where both have results: coding, reasoning and images and charts. Phi 4 Multimodal Instruct wins none.

Which is better for coding?

Qwen3.5 4B. It wins the one coding test both models report.

How do you compare the two?

We use the 14 benchmark tests both models have published scores on. The verdict counts the 9 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 5 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.