Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d271Policy30d66, 14 pendingSources1,015

Gemma 3 12B vs Kimi VL A3B Instruct

Google · released

Wins 1 of 1 area

Images and charts

Moonshot · released

Wins 0 of 1 area

—

Gemma 3 12B is the stronger all-rounder.

Scores updated · 22 tests both models report · How we compare

Where each one wins

Tests won in each of the one area 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.

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.

Gemma 3 12B$0.05 to read · $0.15 to write$0.20Price from deepinfra
Kimi VL A3B InstructNo current price is tracked.

The biggest differences

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

Where Gemma 3 12B pulls ahead

  • College exam questions with charts, maps and diagramsMMMU+2.6points aheadGemma 3 12B59.6Kimi VL A3B Instruct57
  • Harder college exam questions with imagesMMMU-Pro+1.5points aheadGemma 3 12B37.5Kimi VL A3B Instruct36

Where Kimi VL A3B Instruct pulls ahead

  • Math problems shown in pictures and chartsMathVista+5.8points aheadKimi VL A3B Instruct68.7Gemma 3 12B62.9

Every test, side by side

All 22 tests both models report. The winning score is in its model's colour; marks a score checked independently.

Images and chartsGemma 3 12B
  • MathVistaKimi VL A3B Instruct by 5.862.968.7+5.8
  • MMMUGemma 3 12B by 2.659.657+2.6
  • MMMU-ProGemma 3 12B by 1.537.536+1.5

Full images and charts ranking

Other results19 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.

Questions people ask

Which is better, Gemma 3 12B or Kimi VL A3B Instruct?

Gemma 3 12B wins the one area where both have results: images and charts. Kimi VL A3B Instruct wins none.

How do you compare the two?

We use the 22 benchmark tests both models have published scores on. The verdict counts the 3 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 19 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.