Seed 1.8

ByteDanceProprietary

描述

Optimized specifically for multimodal agent scenarios. It features enhanced agent capabilities, upgraded multimodal comprehension, and more flexible context management.

发布日期

2026-02-17

参数规模

—

上下文长度

—

支持模态

image, text, video

能力雷达图

general

100

coding

reasoning

science估算

agents

multimodal

缺少专门科学评测时，Science 由 LLM Stats 科学得分或推理能力估算。

排行榜排名

领域	#排名	分数	来源
智能体能力模型榜	85	37.0	LS
多模态榜	91	48.0	LS

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

AndroidWorld

70.7%自报

Chat

Multi-Challenge

66.7%自报

Coding

AetherCode

38.2%自报

General

MMLU

92.3%自报

Language

MMLU-Pro

84.9%自报

Math

AIME 2025

94.3%自报

MathVista

87.7%自报

MathVision

81.3%自报

Beyond AIME

77.0%自报

DynaMath

61.5%自报

AMO Bench

60.0%自报

Multimodal

MMMU

83.4%自报

VideoMMMU

82.7%自报

OSWorld

61.9%自报

Physics

PHYBench

41.0%自报

Reasoning

LiveCodeBench Pro

1930.00 / 3000自报

BrowseComp-zh

81.3%自报

LiveCodeBench v6

79.5%自报

SWE-Bench Verified

72.9%自报

CharXiv-R

71.4%自报

ARC-AGI

67.9%自报

BrowseCompOpenAI (2025)

67.6%自报

SuperGPQA

64.8%自报

Terminal-Bench 2.0Stanford × Laude Institute (2026)

45.2%自报

Multi-SWE-Bench

42.0%自报

Humanity's Last Exam (with tools, text-only)

40.9%自报

Search

WideSearch

63.8%自报

Tool Calling

BFCL-V4

57.2%自报

Video

LiveSports-3K

77.5%自报

OVOBench

72.6%自报

VideoSimpleQA

67.8%自报

OVBench

65.1%自报

Minerva

62.4%自报

TOMATO

60.8%自报

Vision

CountBench

96.30 / 100自报

SimpleVQA

65.40 / 100自报

RefSpatialBench

56.30 / 100自报

ZEROBench

11.00 / 100自报

VLMsAreBlind

93.0%自报

AI2D

89.1%自报

VideoMME w sub.

87.8%自报

TempCompass

86.9%自报

MMStar

79.9%自报

MuirBench

78.7%自报

LongVideoBench

77.4%自报

BLINK

74.3%自报

MMMU-Pro

73.2%自报

MMVU

73.1%自报

LVBench

73.0%自报

TVBench

71.5%自报

MotionBench

70.6%自报

DUDE

69.4%自报

Hallusion Bench

63.9%自报

VLMsAreBiased

62.0%自报

EMMA

60.9%自报

ERQA

58.8%自报

OmniDocBench 1.5

10.6%自报

AA 评测指数

(Artificial Analysis)

暂无 AA 评测数据

LLM Stats 分类评分

(LLM Stats (zeroeval))

Code

100

Image To Text

Grounding

Reasoning

General

Spatial Reasoning

Vision

Multimodal

Language

Legal

Finance

Healthcare

Chat

Knowledge

Long Context

Math

Frontend Development

Communication

Video

Agents

Chemistry

Economics

Physics

Tool Calling

Science

Coding

Structured Output

定价

输入价格$0 / 1M tokens

输出价格$0 / 1M tokens

混合价格(3:1)$0 / 1M tokens

速度

暂无速度数据

供应商价格排行

3 个供应商

最便宜: ByteDance最贵: Requesty

供应商输入输出

1ByteDance主要

2DeepInfra

3Requesty

$0.25

比较该模型在不同 API 供应商之间的定价。

外部链接

LLM Stats Artificial Analysis