Model index

Model directory

Search the index by provider or use case; open a card for specs, capabilities, and typical applications.

39Models indexed

14Providers

⬡Compare ⇄Matching guide ✦Prompts

BY PROVIDER

Browse results

14 vendors · 39 models

DeepSeek

3 series · 4 models

V4 Series

DeepSeek-V4-Pro

deepseek-v4-pro

V4 flagship — high performance, strong reasoning, very low call cost

Parameters: 1.6T total (49B active per inference)
Context Window: 1M tokens (max output 384K)

DeepSeek-V4-Flash

deepseek-v4-flash

V4 high-speed, low-cost edition — strong reasoning, very low call cost

Parameters: 284B total (13B active per inference)
Context Window: 1M tokens (max output 384K)

DeepSeek-R1

deepseek-r1

Chain-of-thought reasoning — math, coding, complex logic, visible internal thinking

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

DeepSeek-V3 / V3.1

deepseek-v3

High value — MoE, coding, reasoning, long-context optimization

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Alibaba

5 series · 5 models

Qwen3.7-Max

qwen3-7-max

AI Agent — 35h autonomous runs, coding agent, tools, code gen, 1M context

Parameters: Not publicly disclosed
Context Window: 1M tokens

Qwen3-Max / Qwen-Max

qwen3-max

Flagship — strong overall capability, multimodal (image/video/audio), enterprise

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Qwen-Plus / Qwen-Flash

qwen-plus-flash

Plus for value — Flash for speed; high concurrency, low latency, RAG

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Qwen-Long

qwen-long

Ultra-long context (1M tokens) — long-doc summary, filings, legal contracts

Parameters: Not publicly disclosed
Context Window: 1M tokens

Quark

quark-ai

AI search, smart scanning, and learning assistant

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Baidu AI Cloud

3 series · 3 models

ERNIE 5.1

ernie-5-1

May 2026 — elastic training, 1/3 compression, ~6% peer training cost, writing, search, agents

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

ERNIE 4.5 / 4.0 Turbo

ernie-4-5-turbo

High performance, fast inference, multimodal (text-to-image & image understanding), 128K context

Parameters: Not publicly disclosed
Context Window: 128K tokens

ERNIE-X1

ernie-x1

Deep-thinking model for complex reasoning and tool collaboration

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Baichuan AI

1 series · 1 models

Baichuan-M3 Plus

baichuan-m3-plus

Medical vertical AI model for serious healthcare — evidence anchoring, 2.6% hallucination rate

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

01.AI

1 series · 1 models

Yi-Large

yi-large

100B+ parameter closed flagship — AlpacaEval 2.0 on par with GPT-4

Parameters: 100B+ parameters (closed-source flagship)
Context Window: 16K tokens (base version)

ByteDance

3 series · 4 models

Doubao-Seedream 4.5

seedream

Image generation — multi-round edit, reference-image edit & variants

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Seed 1.x

Doubao-Seed-1.8

doubao-seed-1-8

Latest flagship dialogue — knowledge, math logic & programming

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Doubao-Seed-1.6

doubao-seed-1-6

Thinking & Flash — stronger reasoning or low latency for diverse business scenarios

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Jimeng AI

jimeng-ai

AI image generation and AI video generation

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Zhipu AI

4 series · 4 models

GLM-5

glm-5

Iterative upgrade — general foundation capabilities and reasoning efficiency

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

GLM-4.7

glm-4-7

Agentic Coding, long-horizon planning, tools, code, natural dialogue & immersive writing

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

GLM-4V

glm-4v

Multimodal vision understanding and image analysis

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

ChatGLM

chatglm-app

General dialogue, agent tasks, and multimodal understanding

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Moonshot AI

1 series · 2 models

Kimi K2

Kimi K2.6

kimi-k2-6

Latest flagship — swarm mode, large-scale agent collaboration, multimodal, thinking & code

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Kimi K2 Thinking

kimi-k2-thinking

Chain-of-thought — think while using tools, complex reasoning & tool calls

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Tencent Hunyuan

2 series · 2 models

Hy3 Preview

hunyuan-hy3-preview

295B / 21B active MoE — 256K context, deep thinking, agents & coding

Parameters: 295B total / 21B active (MoE)
Context Window: 256K tokens (input 192K / output 128K; ~140K Chinese characters)

Hy2.0

hunyuan-hy2

Previous-generation general model — multimodal text-to-image/video & creative marketing

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

360

2 series · 2 models

360zhinao-turbo-doubao-seed-1-8

360-zhinao-turbo-seed-1-8

32K context — TTFT 0.34s, 47.19 tps, Reasoning, multi-format APIs

Parameters: Not publicly disclosed
Context Window: 32,000 tokens (max output 8,192)

Nomi AI

nano-ai

AI search and intelligent assistant

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Kunlun Tech

3 series · 3 models

SkyClaw-v1.0

skyclaw-v1

Native Agent model — built for execution, optimized for tool calling

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

SkyClaw-v1.0-lite

skyclaw-v1-lite

Lite Agent edition — faster, lower cost, core capabilities retained

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Tiangong AI

tiangong-ai

Intelligent search, AI dialogue, and content generation

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

SenseTime

1 series · 1 models

SenseNova 6.7 Flash-Lite

sensenova-6-7-flash-lite

Lightweight multimodal agent — native vision, 60% lower token use in search

Parameters: Smaller parameter scale (multimodal agent architecture)
Context Window: Not publicly disclosed

JD Technology

2 series · 2 models

48B General Foundation Model (Open Source)

jd-480b-open

First open-source release — 48B params, Apache 2.0, multilingual & multimodal

Parameters: 48 billion parameters
Context Window: Long-context processing (per product description)

Jingyan Enterprise

jd-jingyan-enterprise

Industry-tuned enterprise edition — private deploy, ~50% lower cost vs intl. peers

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Other

1 series · 5 models

General

Spark

iflytek-spark

Voice interaction, multilingual translation, and intelligent Q&A

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Hailuo AI

hailuo-ai

Video generation, content creation, and AI dialogue

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Metaso AI Search

metaso-search

AI-powered search, information refinement, and source attribution

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Kling AI

kling-ai

AI video generation and visual design

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

WPS AI

wps-ai

Smart document writing, PPT generation, and spreadsheet processing

Parameters: Not publicly disclosed
Context Window: Not publicly disclosed

Independent directory. Not affiliated with listed AI providers. All product names and trademarks are the property of their respective owners.

Full legal notice →