Seed 2.1 Turbo

ByteDanceProprietary

描述

ByteDance's low-cost, low-latency next-generation agent model for large-scale production. A deep-thinking model with the full feature set of Seed 2.1 Pro and comparable performance, it is designed for enterprise-grade deployments handling high volumes of online calls across coding, long-chain agents, and multimodal tasks. Served via Volcano Engine as Doubao-Seed-2.1-turbo.

發布日期

2026-06-24

參數規模

—

上下文長度

262K

支援模態

image, text, video

能力雷達圖

general

coding

reasoning

science估算

agents

multimodal

缺少專門科學評測時，Science 由 LLM Stats 科學得分或推理能力估算。

排行榜排名

領域	#排名	分數	來源
智慧體能力模型榜	61	50.0	LS
多模態榜	25	58.0	LS

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

3d

BLINK

79.4%自報

Agents

BrowseCompOpenAI (2025)

84.9%自報

GDPval

82.7%自報

MCP Atlas

80.3%自報

OSWorld

76.4%自報

Web Bench

73.6%自報

OfficeQA Pro

71.1%自報

MobileWorld

70.0%自報

Terminal-Bench 2.1

67.6%自報

CyberGym

67.0%自報

OneMillion Bench

66.6%自報

SeedClawBench

63.8%自報

WildClawBench

62.8%自報

Trae Code Gen

59.7%自報

SWE-Bench ProPrinceton NLP (2024)

57.0%自報

Trae Error Fix

56.7%自報

xDailyBench

56.4%自報

Finance Agent v1.1

56.0%自報

Workspace Bench

54.7%自報

Agent Startup Bench

54.0%自報

Program Bench

49.4%自報

Toolathlon

49.1%自報

Doubao Multi-Turn Bench

49.0%自報

PresentBench

48.3%自報

Repo Env

46.7%自報

ClawEval-MM

46.0%自報

NL2Repo

43.7%自報

CreativeWork

34.5%自報

SWE-Atlas

30.6%自報

APEX-Agents

29.2%自報

GameWorld

25.9%自報

DeepSWE

23.0%自報

PostTrainBench

18.3%自報

Biology

SciCode

57.8%自報

Chemistry

SuperGPQA

67.4%自報

SuperChem

56.6%自報

Code

FrontierCS

50.8%自報

Artifacts Bench

47.0%自報

Image2FloorPlan

35.9%自報

Coding

AetherCode

67.9%自報

Embodied

EmbSpatialBench

0.82 / 100自報

General

MMMU-Pro

82.2%自報

SimpleVQA

0.71 / 100自報

KINA

46.6%自報

MSQA

42.0%自報

Image To Text

OCRBench_V2

62.8%自報

Knowledge

VideoSimpleQA

71.4%自報

WorldBench

63.7%自報

Long Context

DUDE

83.1%自報

LongVideoBench

80.6%自報

MMLongBench-128K

76.9%自報

LVBench

76.8%自報

Math

MathVision

92.7%自報

MathVista

90.5%自報

MathVerse

89.2%自報

Beyond AIME

88.0%自報

EMMA

78.4%自報

FrontierScience Olympiad

76.0%自報

DynaMath

68.1%自報

Humanity's Last Exam

54.6%自報

MathArena Apex

35.4%自報

LiveMathematicianBench

27.7%自報

HorizonMath

2.0%自報

Multimodal

CharXiv-D

94.6%自報

Video-MME

89.0%自報

CharXiv-R

83.6%自報

OVOBench

79.2%自報

TVBench

77.2%自報

LiveSports-3K

77.1%自報

MotionBench

74.8%自報

TreeBench

71.1%自報

ChartQAPro

70.9%自報

OVBench

69.7%自報

VLMsAreBiased

68.3%自報

VideoHolmes

67.6%自報

Minerva

65.9%自報

CrossVid

63.2%自報

BabyVision

62.9%自報

ContPhy

61.1%自報

MeasureBench

58.9%自報

TOMATO

56.8%自報

VisuLogic

0.53 / 100自報

WorldVQA

48.6%自報

VisFactor

43.9%自報

MMSIBench

31.4%自報

ZEROBench

0.11 / 100自報

Reasoning

ERQA

71.3%自報

ArcAGI2

61.3%自報

FrontierScience Research

33.3%自報

Spatial Reasoning

RealWorldQA

86.3%自報

AA 評測指數

(Artificial Analysis)

暫無 AA 評測資料

LLM Stats 分類評分

(LLM Stats (zeroeval))

Structured Output

Legal

Spatial Reasoning

Embodied

Finance

Long Context

Multimodal

Safety

Image To Text

Healthcare

Economics

Tool Calling

Video

Vision

Math

Physics

Reasoning

Knowledge

General

Agents

Biology

Chemistry

Coding

Science

Frontend Development

Code

Systems

定價

輸入價格$0.5 / 1M tokens

輸出價格$2.5 / 1M tokens

混合價格(3:1)$1 / 1M tokens

速度

暫無速度資料

供應商價格排行

4 個供應商

最便宜: ByteDance最貴: Kilo Gateway

供應商輸入輸出

1ByteDance主要

$0.5

$2.5

2NanoGPT

$0.5

$2.5

3OpenRouter

$0.5

$2.5

4Kilo Gateway

$0.5

$2.5

比較該模型在不同 API 供應商之間的定價。

外部連結

LLM Stats Artificial Analysis