模型广场

A natively omni-modal model with unified understanding of text, image, audio, and video, providing a flagship foundation for full-modality capabilities.

上下文

128K

输入 / 1M tokens

$0.84

输出 / 1M tokens

$3.28

缓存读 / 1M

$0

缓存写 / 1M

$0

An efficient iteration of the flagship line, delivering major gains in reasoning, agentic handling, and search with far fewer parameters, balancing high performance and low-cost deployment.

上下文

128K

输入 / 1M tokens

$0.56

输出 / 1M tokens

$2.53

缓存读 / 1M

$0

缓存写 / 1M

$0

Turbo tier in Seed 2.1: multimodal model for coding and long-horizon agents, suited to multi-step task execution and visual content understanding.

上下文

262K

输入 / 1M tokens

$0.5

输出 / 1M tokens

$2.5

缓存读 / 1M

$0.1

缓存写 / 1M

$0.008

Agent-oriented foundation model tuned for tool calling and complex instruction following, with use cases spanning GUI agents, search agents, and multi-step task orchestration.

上下文

256K

输入 / 1M tokens

$0.5

输出 / 1M tokens

$4

缓存读 / 1M

$0.05

缓存写 / 1M

$0.008333

Code-focused preview model for agent workflows, with competitive SWE-Bench and LiveCodeBench scores, aimed at autonomous coding agents and multi-language software tasks.

上下文

256K

输入 / 1M tokens

$0.5

输出 / 1M tokens

$3

缓存读 / 1M

$0.1

缓存写 / 1M

$0.008333

Lightweight tier of the Seed 2.0 family tuned for low-latency agent, coding, and GUI workloads.

上下文

256K

输入 / 1M tokens

$0.25

输出 / 1M tokens

$2

缓存读 / 1M

$0.05

缓存写 / 1M

$0.008333

Flagship of the Seed 2.0 series for complex reasoning and long-horizon agent workflows with reliable tool use.

上下文

256K

输入 / 1M tokens

$0.5

输出 / 1M tokens

$3

缓存读 / 1M

$0.1

缓存写 / 1M

$0.008333

Open-source reasoning model with visible chain-of-thought, tuned for math, coding, and multi-step problem solving.

上下文

164K

输入 / 1M tokens

$0.7

输出 / 1M tokens

$2.5

缓存读 / 1M

$0

缓存写 / 1M

$0

Open-source general-purpose LLM with representative instruction-following and coding skills within the open-weights ecosystem, suited for chat, code assistance, and enterprise text workflows.

上下文

131K

输入 / 1M tokens

$0.287

输出 / 1M tokens

$1.147

缓存读 / 1M

$0.057

缓存写 / 1M

-

Hybrid model that switches between thinking and non-thinking modes on a single endpoint, post-trained for tool use and code agent workflows.

上下文

164K

输入 / 1M tokens

$0.574

输出 / 1M tokens

$1.721

缓存读 / 1M

$0.13

缓存写 / 1M

$0

DeepSeek V4 flagship with adjustable reasoning depth, built for full-codebase analysis, multi-step automation, and complex reasoning tasks.

上下文

1M

输入 / 1M tokens

$0.66

输出 / 1M tokens

$1.98

缓存读 / 1M

$0.022

缓存写 / 1M

$0

A high-performance general-purpose model from Google, designed for advanced reasoning, coding, mathematics, and scientific tasks. Its built-in thinking capabilities improve response accuracy and enable deeper contextual understanding.

上下文

1.1M

输入 / 1M tokens

$0.3

输出 / 1M tokens

$2.5

缓存读 / 1M

$0.03

缓存写 / 1M

$0.08333

Speed-tier variant in the Gemini 3 line, pairing Pro-grade reasoning with Flash-level latency and cost, built for agent workflows and high-throughput interactive apps.

上下文

1M

输入 / 1M tokens

$0.5

输出 / 1M tokens

$3

缓存读 / 1M

$0.05

缓存写 / 1M

$0

A high-efficiency multimodal lite model for low-latency, high-volume workloads like translation, classification, and data extraction, priced at about half of Gemini 3 Flash.

上下文

1M

输入 / 1M tokens

$0.25

输出 / 1M tokens

$1.5

缓存读 / 1M

$0.025

缓存写 / 1M

$0.08333

Cost-effective general model for balanced speed and quality.

上下文

128K

输入 / 1M tokens

$0.31

输出 / 1M tokens

$1.23

缓存读 / 1M

$0.031

缓存写 / 1M

$0.39

Built for conversation and bilingual chat.

上下文

197K

输入 / 1M tokens

$0.304

输出 / 1M tokens

$1.213

缓存读 / 1M

$0.061

缓存写 / 1M

$0

A high-throughput variant of M2.5 with matching quality at roughly triple the speed, tuned for coding and agentic tool use under latency-sensitive workloads.

上下文

200K

输入 / 1M tokens

$0.6

输出 / 1M tokens

$2.4

缓存读 / 1M

$0.03

缓存写 / 1M

$0.375

A next-generation autonomous language model that uses multi-agent collaboration to plan, execute, and continuously refine complex tasks. It supports production-grade workflows including live debugging, root cause analysis, financial modeling, and document generation across Word, Excel, and PowerPoint.

上下文

205K

输入 / 1M tokens

$0.3

输出 / 1M tokens

$1.2

缓存读 / 1M

$0.06

缓存写 / 1M

$0

Speed-optimized variant of M2.7 with identical outputs, tuned for low-latency coding and agent tool-calling workloads.

上下文

200K

输入 / 1M tokens

$0.6

输出 / 1M tokens

$2.4

缓存读 / 1M

$0.06

缓存写 / 1M

$0.375

MiniMax\\'s first multimodal release, accepting image and video input. Long-context inference runs cheaper and faster, and it handles multi-step agent workflows and computer-use scenarios.

上下文

1M

输入 / 1M tokens

$1.2

输出 / 1M tokens

$4.8

缓存读 / 1M

$0.12

缓存写 / 1M

$0

Instruction-tuned for agent workflows with reliable tool calling and multilingual coding.

上下文

131K

输入 / 1M tokens

$0.574

输出 / 1M tokens

$2.3

缓存读 / 1M

$0.115

缓存写 / 1M

$0

Open-source thinking model built for long-horizon reasoning and hundreds of sequential tool calls across autonomous research, coding, and agent workflows.

上下文

262K

输入 / 1M tokens

$0.6

输出 / 1M tokens

$2.5

缓存读 / 1M

$0.15

缓存写 / 1M

$0

Top Chinese-English bilingual. Strong long context.

上下文

262K

输入 / 1M tokens

$0.6

输出 / 1M tokens

$3.011

缓存读 / 1M

$0.15

缓存写 / 1M

$0.718

Latest multimodal model in the K2 series, built for long-horizon coding, code-driven UI/UX generation, and multi-agent orchestration.

上下文

256K

输入 / 1M tokens

$0.8939

输出 / 1M tokens

$3.7131

缓存读 / 1M

$0.34

缓存写 / 1M

$1.1174

Coding-focused variant in the Kimi K2 family. Accepts text and image inputs, with reasoning on by default. Suited for long-horizon coding and agentic workflows.

上下文

262K

输入 / 1M tokens

$0.95

输出 / 1M tokens

$4

缓存读 / 1M

$0.19

缓存写 / 1M

$1.1174

Near GPT-4o performance at lower latency and cost, suited for high-frequency interactions, coding, and vision tasks.

上下文

1M

输入 / 1M tokens

$0.4

输出 / 1M tokens

$1.6

缓存读 / 1M

$0.1

缓存写 / 1M

-

Fast and low-cost; for high-concurrency text and reasoning tasks.

上下文

400K

输入 / 1M tokens

$0.25

输出 / 1M tokens

$2

缓存读 / 1M

$0.025

缓存写 / 1M

$0

Lightweight GPT-5.4 sibling tuned for coding, tool use, and agent workflows at roughly twice the speed of the 5.4 Pro tier.

上下文

400K

输入 / 1M tokens

$0.75

输出 / 1M tokens

$4.5

缓存读 / 1M

$0.075

缓存写 / 1M

$0

Lightweight variant in the gpt-5.4 family, tuned for low-latency, high-volume tasks like classification, extraction, and sub-agent execution.

上下文

400K

输入 / 1M tokens

$0.2

输出 / 1M tokens

$1.25

缓存读 / 1M

$0.02

缓存写 / 1M

$0

Entry point of the GPT-5.6 lineup. A small, low-latency workhorse for chat, tagging, and lightweight agent loops, keeping reasoning solid enough for routine work.

上下文

1.1M

输入 / 1M tokens

$0.2

输出 / 1M tokens

$1.2

缓存读 / 1M

$0.02

缓存写 / 1M

$0.25

Cost-efficient reasoning model with adjustable thinking effort, tuned for STEM and coding tasks where deliberation matters.

上下文

200K

输入 / 1M tokens

$1.1

输出 / 1M tokens

$4.4

缓存读 / 1M

$0.55

缓存写 / 1M

-

Lightweight o-series reasoning model with built-in chain-of-thought, delivering steady math, coding, and tool-use quality at low latency and cost.

上下文

2M

输入 / 1M tokens

$1.1

输出 / 1M tokens

$4.4

缓存读 / 1M

$0.275

缓存写 / 1M

$0

Instruction-tuned Qwen3 variant without thinking mode, strong on multilingual reasoning, math, code, and tool use for agent workflows.

上下文

262K

输入 / 1M tokens

$0.287

输出 / 1M tokens

$2.868

缓存读 / 1M

-

缓存写 / 1M

-

Internal fine-tune of Qwen3-32B for text-only Chinese workloads, tuned to in-house instruction style for QA, summarization and rewriting.

上下文

131K

输入 / 1M tokens

$0.287

输出 / 1M tokens

$1.147

缓存读 / 1M

-

缓存写 / 1M

-

Qwen3-generation thinking-mode reasoning model with native tool use, tuned for math, coding, and multi-step agentic workflows.

上下文

262K

输入 / 1M tokens

$0.359

输出 / 1M tokens

$1.434

缓存读 / 1M

$0.072

缓存写 / 1M

$2.438

Deep-thinking preview that reasons step-by-step before answering, holding up on multi-step math, code, and Chinese-language tasks.

上下文

256K

输入 / 1M tokens

$0.861

输出 / 1M tokens

$3.441

缓存读 / 1M

$0.173

缓存写 / 1M

-

Instruction-tuned non-thinking chat model that answers directly without exposing reasoning, suited to agent workflows needing deterministic output like coding assistance and tool calling.

上下文

262K

输入 / 1M tokens

$0.15

输出 / 1M tokens

$1.2

缓存读 / 1M

-

缓存写 / 1M

-

Plus-tier vision-language model in the Qwen3-VL line, handling text, image and video inputs with strengths in document parsing, video understanding, spatial grounding and agent tool use.

上下文

262K

输入 / 1M tokens

$0.144

输出 / 1M tokens

$1.434

缓存读 / 1M

$0.029

缓存写 / 1M

$0.539

Thinking-mode variant of the dense open-weight model that outputs step-by-step reasoning; at 27B it surpasses the prior open-weight 397B-A17B flagship across coding, math and multi-step reasoning benchmarks.

上下文

262K

输入 / 1M tokens

$0.412564

输出 / 1M tokens

$2.475384

缓存读 / 1M

$0

缓存写 / 1M

$0

Open-weight coding model tuned for agentic terminal tasks and repo-scale reasoning with low inference cost.

上下文

262K

输入 / 1M tokens

$0.248

输出 / 1M tokens

$1.485

缓存读 / 1M

-

缓存写 / 1M

-

Enhanced Qwen model with strong bilingual (Chinese/English) comprehension, excelling at long-document analysis and structured output.

上下文

1M

输入 / 1M tokens

$0.276

输出 / 1M tokens

$1.651

缓存读 / 1M

$0

缓存写 / 1M

$0

Flagship agent-centric reasoning model with a 1M context window, excelling at coding, productivity, and long-horizon autonomous tasks.

上下文

1M

输入 / 1M tokens

$1.65

输出 / 1M tokens

$4.951

缓存读 / 1M

$0.33

缓存写 / 1M

$2.063

Qwen 3.7 Plus is a mid-tier multimodal model that reads screens, operates GUIs, and navigates mobile apps end-to-end. Suited for agent workflows and tool calling.

上下文

1M

输入 / 1M tokens

$0.276

输出 / 1M tokens

$1.101

缓存读 / 1M

$0.056

缓存写 / 1M

$0

Text-only model in the GLM-4 family with improved coding benchmarks and tool calls during reasoning, suited for software development and agent workflows.

上下文

203K

输入 / 1M tokens

$0.574

输出 / 1M tokens

$2.294

缓存读 / 1M

$0.115

缓存写 / 1M

$0

Balanced model with strong Chinese understanding.

上下文

128K

输入 / 1M tokens

$0.59

输出 / 1M tokens

$2.35

缓存读 / 1M

$0.55

缓存写 / 1M

$0

Competitive general-purpose Chinese model.

上下文

203K

输入 / 1M tokens

$0.88

输出 / 1M tokens

$3.23

缓存读 / 1M

$0.172

缓存写 / 1M

$0

Designed for agent-based workflows such as OpenClaw.

上下文

128K

输入 / 1M tokens

$1.2

输出 / 1M tokens

$4

缓存读 / 1M

$0.24

缓存写 / 1M

$0

Offers significantly improved coding and long-horizon task capabilities, autonomously planning, executing, and refining a single task for over eight hours to deliver complete, engineering-grade results.

上下文

128K

输入 / 1M tokens

$0.826

输出 / 1M tokens

$3.303

缓存读 / 1M

$1.035

缓存写 / 1M

$1.376

GLM 5.2 is a large-scale reasoning model with a 1M-token context window, suited for long-horizon agents, repo-level coding, and multi-step automation.

上下文

1M

输入 / 1M tokens

$1.4

输出 / 1M tokens

$4.4

缓存读 / 1M

$0.275

缓存写 / 1M

$0

Flagship-tier model for complex software engineering and long-horizon agent tasks, with stable execution across multi-turn tool calls and large-scale codebase refactoring.

上下文

1M

输入 / 1M tokens

$1.4

输出 / 1M tokens

$4.4

缓存读 / 1M

$0.26

缓存写 / 1M

$0