Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

Seed 1.8 vs Seed 2.1 Pro Preview

22 SHARED BENCHMARKS

Across 22 shared benchmarks, Seed 1.8 scores higher on 0 and Seed 2.1 Pro Preview on 22. The widest gap is ZeroBench, where Seed 2.1 Pro Preview scores 18 against 11.

BYTEDANCEVSBYTEDANCE22 SHARED0–22 HEAD-TO-HEADNEWEST SCORE ADDED

At a glance

MakerByteDanceByteDance
Released——
Price per 1M tokens input / output$0.25 / $2.00—
Cost of 1M in + 1M out$2.25—
Head-to-head of 22 shared benchmarks0 wins22 wins
Scores tracked independently verified40 0 ◆66 1 ◆

Prices: deepinfra for Seed 1.8. ◆ = independently verified score.

Where each leadsby capability area · benchmark wins

AreaSeed 1.8WINSSeed 2.1 Pro Preview
Agentic02Seed 2.1 Pro Preview leads 2 of 2 · widest: BrowseComp 67.6 vs 86.2
Multimodal04Seed 2.1 Pro Preview leads 4 of 4 · widest: CharXiv (RQ) 71.4 vs 86.4

Biggest gaps

Seed 1.8 pulls furthest ahead on

No ratified-area lead of 3 points or more.

Seed 2.1 Pro Preview pulls furthest ahead on

  1. BrowseComp86.2 vs 67.6
  2. OSWorld-Verified78.8 vs 61.9
  3. CharXiv (RQ)86.4 vs 71.4

Every shared benchmark22 · grouped by area

Agentic 2

Multimodal 4

BLINK74.381.4
CharXiv (RQ)71.486.4
MathVista87.790.7
MMMU-Pro73.282.7

Other shared benchmarks 16

DUDE69.482.8
DynaMath61.573.1
ERQA58.872
MATH-Vision81.394.5
Minerva62.470.7
MotionBench70.674.9
OVBench65.170
SimpleVQA65.474.5
SuperGPQA64.870.8
TOMATO60.879.5
TVBench71.580.5

Each score is the one the model's own page shows — the most authoritative tracked result for that benchmark, on the benchmark's standard methodology (source-attributed). ↓ marks lower-is-better metrics; ◆ an independently verified score. Open either model for its full surface, provenance and pricing.

Compare Seed 2.1 Pro Preview withALL PAIRINGS →