Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

Seed 1.8 vs Seed 2.1 Pro Preview

ByteDance

Wins 0 of 2 areas

—

ByteDance

Wins 2 of 2 areas

Agents · Images and charts

Seed 2.1 Pro Preview 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 two 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.

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.

Seed 1.8$0.25 to read · $2.00 to write$2.25Price from deepinfra
Seed 2.1 Pro PreviewNo 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 Seed 2.1 Pro Preview pulls ahead

  • Finds hard-to-locate facts by browsing the webBrowseComp+18.6points aheadSeed 2.1 Pro Preview86.2Seed 1.867.6
  • Completes tasks by operating a computer desktopOSWorld-Verified+16.9points aheadSeed 2.1 Pro Preview78.8Seed 1.861.9
  • Reasoning about charts from research papersCharXiv (RQ)+15points aheadSeed 2.1 Pro Preview86.4Seed 1.871.4

Where Seed 1.8 pulls ahead

No clear win on a test scored out of 100.

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.

AgentsSeed 2.1 Pro Preview

Full agents ranking

Images and chartsSeed 2.1 Pro Preview
  • CharXiv (RQ)Seed 2.1 Pro Preview by 1571.486.4+15
  • MMMU-ProSeed 2.1 Pro Preview by 9.573.282.7+9.5
  • BLINKSeed 2.1 Pro Preview by 7.174.381.4+7.1
  • MathVistaSeed 2.1 Pro Preview by 387.790.7+3

Full images and charts ranking

Other results16 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.

  • TOMATOSeed 2.1 Pro Preview by 18.760.879.5+18.7
  • DUDESeed 2.1 Pro Preview by 13.469.482.8+13.4
  • ERQASeed 2.1 Pro Preview by 13.258.872+13.2
  • MATH-VisionSeed 2.1 Pro Preview by 13.281.394.5+13.2
  • DynaMathSeed 2.1 Pro Preview by 11.661.573.1+11.6
  • BeyondAIMESeed 2.1 Pro Preview by 107787+10
  • SimpleVQASeed 2.1 Pro Preview by 9.165.474.5+9.1
  • TVBenchSeed 2.1 Pro Preview by 971.580.5+9
  • VideoSimpleQASeed 2.1 Pro Preview by 8.667.876.4+8.6
  • MinervaSeed 2.1 Pro Preview by 8.362.470.7+8.3
  • ZeroBenchSeed 2.1 Pro Preview by 71118+7
  • SuperGPQASeed 2.1 Pro Preview by 664.870.8+6
  • LVBenchSeed 2.1 Pro Preview by 57378+5
  • OVBenchSeed 2.1 Pro Preview by 4.965.170+4.9
  • MotionBenchSeed 2.1 Pro Preview by 4.370.674.9+4.3
  • LongVideoBenchSeed 2.1 Pro Preview by 3.277.480.6+3.2

Questions people ask

Which is better, Seed 1.8 or Seed 2.1 Pro Preview?

Seed 2.1 Pro Preview wins all two areas where both have results: agents and images and charts. Seed 1.8 wins none.

How do you compare the two?

We use the 22 benchmark tests both models have published scores on. The verdict counts the 6 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 16 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.