Seed 2.0 Mini

Name: Seed 2.0 Mini
Author: ByteDance

Multimodal

ByteDance

Built for low-latency, high-concurrency, cost-sensitive use cases, with flexible deployment, four-tier thinking, and multimodal

Key Specifications

Parameters

Context

256.0K

Release Date

February 17, 2026

Average Score

58.5%

API Documentation

Timeline

Key dates in the model's history

Announcement

February 17, 2026

Last Update

September 10, 2026

Today

September 19, 2026

Technical Specifications

Parameters

Training Tokens

Knowledge Cutoff

Family

Capabilities

MultimodalZeroEval

Pricing & Availability

Input (per 1M tokens)

$0.10

Output (per 1M tokens)

$0.40

Max Input Tokens

256.0K

Max Output Tokens

256.0K

Supported Features

Function CallingStructured OutputCode ExecutionWeb SearchBatch InferenceFine-tuning

Benchmark Results

Model performance metrics across various tests and benchmarks

Reasoning

Logical reasoning and analysis

GPQA

Model Card Table 4 efficient-models; GPQA Diamond percentage score mapped to existing GPQA convention. • Self-reported

79.0%

Multimodal

Working with images and visual data

MMMU

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

79.7%

Other Tests

Specialized benchmarks

AetherCode

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

29.8%

AIME 2025

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

87.0%

AIME 2026

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

86.7%

ARC-AGI

Model Card Table 8; ArcAGI1-Image percentage score. • Self-reported

29.8%

ArcAGI2

Model Card Table 4 efficient-models; ARC-AGI-2 text/general-reasoning percentage score. • Self-reported

2.3%

ARC-AGI v2

Model Card Table 8; ArcAGI2-Image percentage score. • Self-reported

1.5%

BabyVision

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

38.7%

Beyond AIME

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

69.0%

BLINK

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

73.4%

ChartQAPro

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

65.2%

CharXiv-D

Model Card Table 8; CharXiv-DQ percentage score. • Self-reported

91.9%

CharXiv-R

Model Card Table 8; CharXiv-RQ percentage score. • Self-reported

70.8%

CL-bench

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

14.8%

CodeForces

Model Card Table 4 efficient-models; Codeforces Elo rating. • Self-reported

54.8%

COLLIE

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

91.2%

ContPhy

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

55.9%

CountBench

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

95.5%

CrossVid

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

58.6%

DUDE

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

68.9%

DynaMath

Model Card Table 8; DynaMath worst-case accuracy over 10 variants. • Self-reported

58.9%

EMMA

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

57.0%

ERQA

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

56.3%

FActScore

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

50.4%

FRAMES

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

80.5%

FrontierScience Olympiad

Model Card Table 4 efficient-models; FrontierScience olympiad split percentage score. • Self-reported

44.0%

FrontierScience Research

Model Card Table 4 efficient-models; FrontierScience research split percentage score. • Self-reported

3.3%

Global PIQA

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

89.2%

Graphwalks BFS <128k

Model Card Table 4 efficient-models; Graphwalks BFS <128K percentage score. • Self-reported

64.1%

Graphwalks parents <128k

Model Card Table 4 efficient-models; Graphwalks Parents <128K percentage score. • Self-reported

93.0%

Hallusion Bench

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

65.1%

HealthBench

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

30.0%

HealthBench Hard

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

15.3%

HiPhO

Model Card Table 8; HiPhO average normalized score over 13 physics olympiad competitions. • Self-reported

55.8%

Humanity's Last Exam (no tools, text-only)

Model Card Table 4 efficient-models; no-tools, text-only HLE percentage score. • Self-reported

13.3%

LiveCodeBench v6

Model Card Table 4 efficient-models; LiveCodeBench v6 percentage score. • Self-reported

64.1%

LiveSports-3K

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

73.3%

LongFact Concepts

Model Card Table 4 efficient-models; LongFact Concepts percentage score. • Self-reported

91.4%

LongFact Objects

Model Card Table 4 efficient-models; LongFact Objects percentage score. • Self-reported

87.0%

LongVideoBench

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

74.8%

LVBench

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

66.6%

MathArena Apex

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

4.2%

MathVision

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

78.1%

MathVista-Mini

Model Card Table 8; MathVista testmini Pass@1 percentage score. • Self-reported

85.5%

Minerva

Model Card Table 9 public video benchmarks; subtitle-marked Minerva score. • Self-reported

54.7%

MMLongBench-Doc

Model Card Table 8; MMLongBench-Doc percentage score. • Self-reported

48.9%

MMLU-Pro

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

83.6%

MMMLU

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

81.6%

MMMU-Pro

Model Card Table 8; MMMU-Pro aggregate over Standard 10-option and Vision subsets. • Self-reported

71.4%

MMStar

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

79.1%

MMVU

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

69.0%

MotionBench

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

64.4%

MRCR v2 (8-needle)

Model Card Table 4 efficient-models; MRCR v2 8-needle percentage score. • Self-reported

51.4%

MTVQA

Model Card Table 8; MTVQA scored by DeepSeek-V3-0324 LLM judge. • Self-reported

50.6%

MuirBench

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

78.0%

Multi-Challenge

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

61.1%

OCRBench_V2

Model Card Table 8; OCRBenchv2 average of overall English and Chinese scores. • Self-reported

58.5%

OmniDocBench 1.5

Model Card Table 8; OmniDocBench 1.5 overall edit distance. • Self-reported

11.0%

OVBench

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

60.1%

OVOBench

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

70.4%

PHYBench

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

56.0%

RealWorldQA

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

81.6%

RefSpatialBench

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

55.6%

SimpleQA Verified

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

18.9%

SimpleVQA

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

68.7%

SuperChem

Model Card Table 4 efficient-models; SuperChem text-only percentage score. • Self-reported

16.2%

SuperGPQA

Model Card Table 4 efficient-models; reported percentage score. • Self-reported

61.6%

TempCompass

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

83.7%

TOMATO

Model Card Table 9 public video benchmarks; default TOMATO score. • Self-reported

47.4%

TreeBench

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

57.3%

TVBench

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

70.5%

VideoHolmes

Model Card Table 9 public video benchmarks; subtitle-marked VideoHolmes score. • Self-reported

58.6%

VideoMME w sub.

Model Card Table 9 public video benchmarks; VideoMME with subtitles. • Self-reported

81.2%

VideoMMMU

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

80.6%

VideoSimpleQA

Model Card Table 9 public video benchmarks; reported percentage score. • Self-reported

67.7%

VisFactor

Model Card Table 8; VisFactor macro-average percentage score. • Self-reported

23.6%

VisuLogic

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

40.4%

VLMsAreBiased

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

58.4%

VLMsAreBlind

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

93.1%

WorldVQA

Model Card Table 8 public visual-language benchmarks; Pass@1 percentage score. • Self-reported

47.6%

ZEROBench

Model Card Table 8; ZEROBench main split percentage score. • Self-reported

7.0%

ZEROBench-Sub

Model Card Table 8; ZEROBench sub split percentage score. • Self-reported

36.2%

License & Metadata

License

proprietary

Announcement Date

February 17, 2026

Last Updated

September 10, 2026

Similar Models

All Models

Seed 2.0 Lite

ByteDance

Best score:0.9 (GPQA)

Released:Feb 2026

Seed 2.1 Pro

ByteDance

Released:Jun 2026

Seed 2.1 Turbo

ByteDance

Released:Jun 2026

Seed 1.8

ByteDance

Best score:0.9 (MMLU)

Released:Feb 2026

Price:$0.25/1M tokens

Seedance 1.5 Pro

ByteDance

Released:Jan 1970

Seed 2.0 Pro

ByteDance

Best score:0.9 (GPQA)

Released:Feb 2026

Price:$0.50/1M tokens

ERNIE 5.0

Baidu

Best score:0.8 (GPQA)

Released:Jan 2025

Grok-4

xAI

Best score:0.9 (GPQA)

Released:Jul 2025

Price:$3.00/1M tokens

Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.