Discover Global AI Models for the Next Generation of Apps

Doubao Seed 1.6 Flash is an image understanding model from Bytedance, designed for vision understanding and document analysis.

Doubao Seedance 1.0 Pro is an video generation model from Bytedance, designed for video generation and motion content production.

Doubao Seed Translation is an text generation model from Bytedance, designed for general AI application workflows.

Doubao 1.5 Thinking Pro is an reasoning model model from Bytedance, designed for reasoning, planning, and complex problem solving.

Doubao 1.5 UI TARS is an reasoning model model from Bytedance, designed for reasoning, planning, and complex problem solving.

Doubao Seed 1.6 Vision is an image understanding model from Bytedance, designed for vision understanding and document analysis.

Native audio-visual full-duplex LLM released August 5, 2026 that listens while speaking: fuses audio, video and text in one end-to-end architecture with in-model turn-taking instead of external voice-activity pipelines.

Video generation model introduced July 31, 2026 around one-take creation: 30-second continuous scenes with setup, development and resolution in a single generation, native 4K with audio, multi-round extension and up to 50 reference assets across 10+ languages.

Doubao Seed 1.6 Thinking is an image understanding model from Bytedance, designed for vision understanding and document analysis.

Doubao 1.5 Thinking Vision Pro is an image understanding model from Bytedance, designed for vision understanding and document analysis.

Doubao 1.5 Vision Lite is an image understanding model from Bytedance, designed for vision understanding and document analysis.

Doubao 1.5 Vision Pro 32k is an image understanding model from Bytedance, designed for vision understanding and document analysis.

Doubao 1.5 Lite 32k is an text generation model from Bytedance, designed for general AI application workflows.

Doubao Embedding is an embedding model from Bytedance, designed for general AI application workflows.

Cinema-grade audio creation model released July 20, 2026 that jointly generates voice, sound effects and ambient sound: decomposes creative intent into a structured timeline, supports ~2-minute generations with continuation and multi-role performance from a single voice.

Efficient image tier of the Seedream 5.0 generation (July 2026).

Image flagship released July 8, 2026 with design-aware generation; 5.0 Flash alongside.

Audio-video joint generation with multi-camera control; international release March 26, 2026.

Doubao Seed 1.6 is an image understanding model from Bytedance, designed for vision understanding and document analysis.

Doubao Seedream 4.0 is an image generation model from Bytedance, designed for image generation and creative visual workflows.

Doubao Seedream 3.0 T2i is an image generation model from Bytedance, designed for image generation and creative visual workflows.

Doubao SeedEdit 3.0 I2i is an image generation model from Bytedance, designed for image generation and creative visual workflows.

Doubao 1.5 Pro 32k is an text generation model from Bytedance, designed for general AI application workflows.

Doubao 1.5 Pro 256k is an text generation model from bytedance, designed for general AI application workflows.

Doubao 1.5 Vision Pro is an image understanding model from bytedance, designed for vision understanding and document analysis.

Doubao Seedance 1.0 Lite I2v is an video generation model from bytedance, designed for video generation and motion content production.

Doubao Seedance 1.0 Lite T2v is an video generation model from bytedance, designed for video generation and motion content production.

Distilled text-to-image variant of Qwen-Image 2.1 released October 9, 2026: 4-6x faster generation at 768x768 native resolution with comparable visual quality, paired with an equally sped-up editing companion.

7B open image model released September 20, 2026; top of Hugging Face text-to-image trending.

Qwen Vl Plus is an image understanding model from Alibaba, designed for vision understanding and document analysis.

Qwen3 Asr Flash is an speech recognition model from Alibaba, designed for speech recognition and transcription.

Qwen Image Plus is an image generation model from Alibaba, designed for image generation and creative visual workflows.

Qwen3 Coder Plus is an text generation model from Alibaba, designed for general AI application workflows.

Qwen3 Vl Plus is an image understanding model from Alibaba, designed for vision understanding and document analysis.

Qwen3 Max is an text generation model from Alibaba, designed for general AI application workflows.

Qwen3 Vl 235b A22b Thinking is an image understanding model from Alibaba, designed for vision understanding and document analysis.

Wan2.5 I2i Preview is an image generation model from Alibaba, designed for image generation and creative visual workflows.

Qwen Image Edit is an image generation model from Alibaba, designed for image generation and creative visual workflows.

Qwen3 Livetranslate Flaltimeash Re 2025 09 22 is an speech recognition model from Alibaba, designed for speech recognition and transcription.

Qwen3 Coder Flash is an text generation model from Alibaba, designed for general AI application workflows.

Qwen Flash is an text generation model from Alibaba, designed for general AI application workflows.

Qwen Mt Turbo is an text generation model from Alibaba, designed for general AI application workflows.

Qwen Mt Plus is an text generation model from Alibaba, designed for general AI application workflows.

Qwen Tts Realtime is an text-to-speech model from Alibaba, designed for speech synthesis and voice experiences.

Qwen Mt Image is an image generation model from Alibaba, designed for image generation and creative visual workflows.

Qwen Deep Research is an text generation model from Alibaba, designed for general AI application workflows.

Qwen3 1.7B is an text generation model from Alibaba, designed for general AI application workflows.

Open-weights architecture preview for Qwen4 released August 24, 2026: 125B MoE (6B activated) with Gated DeltaNet and quadratic-sublinear attention plus a compressed n-gram memory, decoding ~2.5x faster than Qwen3.8-27B at comparable quality.

Next-generation Tongyi Wanxiang video model in public beta since August 2026 on Alibaba Cloud Model Studio: native 30-second 1080p video (up to 4K) with audio generated in one pass, accepting text, image, audio, video, document and web-page input.

First open-weight Max-class Qwen model; current Qwen flagship.

Qwen3 235B A22B Instruct 2507 is an text generation model from Alibaba, designed for general AI application workflows.

Qwen Vl Max is an image understanding model from Alibaba, designed for vision understanding and document analysis.

Qwen Image is an image generation model from Alibaba, designed for image generation and creative visual workflows.

Qwen3 30B A3B Instruct 2507 is an text generation model from Alibaba, designed for general AI application workflows.

Qwen3 0.6B is an text generation model from Alibaba, designed for general AI application workflows.

Wan2.1 T2V 1.3B is an video generation model from Alibaba, designed for video generation and motion content production.

Qwen2.5 VL 32B Instruct is an image understanding model from Alibaba, designed for vision understanding and document analysis.

Qwen3 30B A3B is an text generation model from Alibaba, designed for general AI application workflows.

Qwen3 14B is an text generation model from Alibaba, designed for general AI application workflows.

Wan2.1 I2V 14B 480P is an video generation model from Alibaba, designed for video generation and motion content production.

Qwen3 235B A22B is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 7b Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwen3 4B is an text generation model from Alibaba, designed for general AI application workflows.

Prior Max-class flagship generation.

Released April 2, 2026.

Open-source 397B MoE released in early 2026; trending on Hugging Face.

Latest Qwen coder generation, released February 4, 2026.

Reasoning-enabled Max tier released January 26, 2026.

Qwen3 Next 80B A3B Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Wan2.5 T2v Preview is an video generation model from Alibaba, designed for video generation and motion content production.

Wan2.5 T2i Preview is an image generation model from Alibaba, designed for image generation and creative visual workflows.

Wan2.5 I2v Preview is an video generation model from Alibaba, designed for video generation and motion content production.

Qwen3 Omni Flash Realtime is an full modal model from Alibaba, designed for general AI application workflows.

Qwen3 Omni 30b A3b Captioner is an speech recognition model from Alibaba, designed for speech recognition and transcription.

Qwen3 Tts Flash is an text-to-speech model from Alibaba, designed for speech synthesis and voice experiences.

Qwen3 Tts Flash Realtime is an text-to-speech model from Alibaba, designed for speech synthesis and voice experiences.

API video release (December 16, 2025) with native multimodality, 150 reference frames and multi-speaker dialogue.

Qwen Omni Turbo Realtime is an full modal model from Alibaba, designed for general AI application workflows.

Qwen Tts is an text-to-speech model from Alibaba, designed for speech synthesis and voice experiences.

Qwen Plus is an text generation model from Alibaba, designed for general AI application workflows.

Qwen2 72B Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwq 32B is an text generation model from Alibaba, designed for general AI application workflows.

Qwq Plus is an reasoning model model from Alibaba, designed for reasoning, planning, and complex problem solving.

Qwen3 8B is an text generation model from Alibaba, designed for general AI application workflows.

Qwen3 32B is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 0.5b Base is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 3b Base is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 1.5b Base is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 0.5b Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2 0.5b Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2 1.5b Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 1.5b Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 3b Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwen Max is an text generation model from Alibaba, designed for general AI application workflows.

Qwen Omni Turbo is an full modal model from Alibaba, designed for general AI application workflows.

Qwen V2.5 14b Instruct is an text generation model from Alibaba, designed for general AI application workflows.

Qwen V2.5 7b Base is an text generation model from Alibaba, designed for general AI application workflows.

GA October 6, 2026; the new default image model in the Gemini app and AI Mode.

Google's top image model, generally available October 6, 2026.

Google's frontier model announced September 30, 2026, rolling out in staged previews; text arena leader in the October 2026 snapshot.

Gemini 2.0 Flash Lite is an text generation model from Google, designed for general AI application workflows.

Gemini 2.5 Flash is an text generation model from Google, designed for general AI application workflows.

Gemini 2.5 Flash Lite is an text generation model from Google, designed for general AI application workflows.

Current Flash tier (September 2026) tuned for software engineering and agents; introductory $0.75/$3.75 per 1M tokens through December 31, 2026.

Flash release of August 2026.

Gemma 3n E2B Instructed is an text generation model from Google, designed for general AI application workflows.

Gemma 3n E2B is an text generation model from Google, designed for general AI application workflows.

Gemma 3n E4B is an text generation model from Google, designed for general AI application workflows.

Gemma 3n E2B Instructed LiteRT (Preview) is an text generation model from Google, designed for general AI application workflows.

Gemma 3n E4B Instructed is an text generation model from Google, designed for general AI application workflows.

Gemma 3n E4B Instructed LiteRT Preview is an text generation model from Google, designed for general AI application workflows.

Gemini 2.0 Flash is an text generation model from Google, designed for general AI application workflows.

Gemini 2.5 Pro is an text generation model from Google, designed for general AI application workflows.

Gemini 2.5 Pro Preview 06 05 is an text generation model from Google, designed for general AI application workflows.

Gemini Diffusion is an text generation model from Google, designed for general AI application workflows.

Gemma 3 4B is an text generation model from Google, designed for general AI application workflows.

Gemma 3 1B is an text generation model from Google, designed for general AI application workflows.

Gemma 3 12B is an text generation model from Google, designed for general AI application workflows.

Gemma 3 27B is an text generation model from Google, designed for general AI application workflows.

Gemini 1.5 Flash 8B is an text generation model from Google, designed for general AI application workflows.

Gemini 2.0 Flash Thinking is an text generation model from Google, designed for general AI application workflows.

Gemma 2 9B is an text generation model from Google, designed for general AI application workflows.

Gemma 2 27B is an text generation model from Google, designed for general AI application workflows.

Gemini 1.5 Flash is an text generation model from Google, designed for general AI application workflows.

Gemini 1.5 Pro is an text generation model from Google, designed for general AI application workflows.

Gemini 1.0 Pro is an text generation model from Google, designed for general AI application workflows.

Preview released February 19, 2026; remains a top general-purpose tier.

Current video generation family in Quality, Fast and Lite tiers; Veo 3.0 model IDs retired June 30, 2026.

Newest efficient tier of the GPT-6 family, released September 29, 2026.

Efficient GPT-6 tier released September 22, 2026 at $4/$20 per 1M tokens - roughly 2.5x cheaper than Astra for everyday workloads.

GPT 4.1 Mini is an text generation model from Openai, designed for general AI application workflows.

GPT 5 is an text generation model from Openai, designed for general AI application workflows.

GPT OSS 120B is an text generation model from Openai, designed for general AI application workflows.

OpenAI flagship released September 3, 2026: ~1.05M-token context, 128K output and 5 reasoning levels at $10/$50 per 1M tokens. Best for complex agentic tasks.

GPT 4.5 is an text generation model from Openai, designed for general AI application workflows.

O3 Mini is an text generation model from Openai, designed for general AI application workflows.

GPT 5 Codex is an text generation model from Openai, designed for general AI application workflows.

GPT 5 Mini is an text generation model from Openai, designed for general AI application workflows.

GPT 5 Nano is an text generation model from Openai, designed for general AI application workflows.

Gpt Oss 20b is an text generation model from Openai, designed for general AI application workflows.

O3 is an text generation model from Openai, designed for general AI application workflows.

O4 Mini is an text generation model from Openai, designed for general AI application workflows.

GPT 4.1 is an text generation model from Openai, designed for general AI application workflows.

GPT 4.1 Nano is an text generation model from Openai, designed for general AI application workflows.

GPT 4o is an text generation model from Openai, designed for general AI application workflows.

O1 Pro is an text generation model from Openai, designed for general AI application workflows.

O1 is an text generation model from Openai, designed for general AI application workflows.

Mid-generation refresh shipped July 9, 2026 alongside Terra and Luna; the official replacement path for deprecated GPT-5 snapshots.

O1 Mini is an text generation model from Openai, designed for general AI application workflows.

O1 Preview is an text generation model from Openai, designed for general AI application workflows.

GPT 3.5 Turbo is an text generation model from Openai, designed for general AI application workflows.

GPT 4o Mini is an text generation model from Openai, designed for general AI application workflows.

GPT 4 Turbo is an text generation model from Openai, designed for general AI application workflows.

Image generation model released April 21, 2026 with native visual reasoning, 2K output and multi-image input.

GPT 4 is an text generation model from Openai, designed for general AI application workflows.

Coding flagship of the GPT-5.x line, released February 5, 2026.

Open-weight multimodal tier released September 10, 2026: 552B-parameter MoE with a causal encoder-decoder architecture, native vision input, 1M-token context and 384K max output. API pricing $0.30/$1.20 per 1M tokens at peak (50% off-peak).

Deepseek R1 is an text generation model from Deepseek, designed for general AI application workflows.

Deepseek V3.1 is an reasoning model model from Deepseek, designed for reasoning, planning, and complex problem solving.

Deepseek V3 is an text generation model from Deepseek, designed for general AI application workflows.

Deepseek R1 Distill Qwen 32B is an text generation model from Deepseek, designed for general AI application workflows.

Deepseek R1 Distill Qwen 7B is an text generation model from Deepseek, designed for general AI application workflows.

Deepseek R1 Distill Qwen 14B is an text generation model from Deepseek, designed for general AI application workflows.

Deepseek R1 Distill Llama 8B is an text generation model from Deepseek, designed for general AI application workflows.

Deepseek R1 Distill Qwen 1.5B is an text generation model from Deepseek, designed for general AI application workflows.

284B efficient tier launched April 24, 2026 alongside V4 Pro. Retired from the API on September 10, 2026; superseded by V4.1 Flash.

1.6T-parameter MoE; API launched April 24, 2026 and reached general availability August 13, 2026.

Fast tier released October 7, 2026 at $0.10/$0.50 per 1M tokens for prompts up to 100K (5x beyond) - up to 90% cheaper than Haiku 4.5.

Balanced tier released September 28, 2026 at $2/$10; prompt cache reads halved to $0.10/1M in October 2026.

Anthropic flagship released September 22, 2026 at $4/$20 per 1M tokens with $0.20 cache reads - 60% cheaper than Opus 5.

Claude Haiku 4.5 is an text generation model from Anthropic, designed for general AI application workflows.

Claude Sonnet 4.5 is an text generation model from Anthropic, designed for general AI application workflows.

Claude 3 Sonnet is an text generation model from Anthropic, designed for general AI application workflows.

Claude Opus 4.1 is an text generation model from Anthropic, designed for general AI application workflows.

Claude 3.7 Sonnet is an text generation model from Anthropic, designed for general AI application workflows.

Claude 3 Opus is an text generation model from Anthropic, designed for general AI application workflows.

Claude Sonnet 4 is an text generation model from Anthropic, designed for general AI application workflows.

Claude Opus 4 is an text generation model from Anthropic, designed for general AI application workflows.

Flagship released July 24, 2026 at $5/$25 per 1M tokens.

Claude 3 Haiku is an text generation model from Anthropic, designed for general AI application workflows.

Claude 3.5 Haiku is an text generation model from Anthropic, designed for general AI application workflows.

Released June 30, 2026; succeeded by Sonnet 5.5 in September.

Claude 3.5 Sonnet is an text generation model from Anthropic, designed for general AI application workflows.

Released April 16, 2026; long-horizon coding and agent workloads.

Qianfan Lightning is an text generation model from Baidu, designed for general AI application workflows.

Qianfan VL 8B is an image understanding model from Baidu, designed for vision understanding and document analysis.

Qianfan VL 70B is an image understanding model from Baidu, designed for vision understanding and document analysis.

ERNIE X1 Turbo is an reasoning model model from Baidu, designed for reasoning, planning, and complex problem solving.

Baidu Turbo 2.0 Audio Vision is an video generation model from Baidu, designed for video generation and motion content production.

Qianfan Llama VL 8B is an image understanding model from Baidu, designed for vision understanding and document analysis.

Qianfan ToyTalk is an text generation model from Baidu, designed for general AI application workflows.

Qianfanhuijin 8B is an text generation model from Baidu, designed for general AI application workflows.

Qianfan QI VL is an image understanding model from Baidu, designed for vision understanding and document analysis.

Qianfanhuijin Reason 8B is an text generation model from Baidu, designed for general AI application workflows.

Qianfan PublicOpinion Classification is an text classification model from Baidu, designed for general AI application workflows.

Qianfanhuijin Reason 70B is an text generation model from Baidu, designed for general AI application workflows.

Qianfanhuijin 70B is an text generation model from Baidu, designed for general AI application workflows.

ERNIE X1.1 Preview is an reasoning model model from Baidu, designed for reasoning, planning, and complex problem solving.

ERNIE 4.5 Turbo VL is an full modal model from Baidu, designed for general AI application workflows.

ERNIE 4.5 Turbo is an text generation model from Baidu, designed for general AI application workflows.

Baidu Turbo 2.0 is an video generation model from Baidu, designed for video generation and motion content production.

ERNIE 4.5 21B A3B Thinking is an reasoning model model from Baidu, designed for reasoning, planning, and complex problem solving.

ERNIE 4.5 300B A47B Paddle is an text generation model from Baidu, designed for general AI application workflows.

ERNIE 4.5 21B A3B Paddle is an text generation model from Baidu, designed for general AI application workflows.

ERNIE 4.5 VL 424B A47B Paddle is an full modal model from Baidu, designed for general AI application workflows.

ERNIE 4.5 VL 28B A3B is an full modal model from Baidu, designed for general AI application workflows.

ERNIE 4.5 0.3B is an text generation model from Baidu, designed for general AI application workflows.

PP UIE 7B is an text classification model from Baidu, designed for general AI application workflows.

PP UIE 14B is an text classification model from Baidu, designed for general AI application workflows.

Paddlemix PP DocBee is an full modal model from Baidu, designed for general AI application workflows.

Released May 8, 2026: lightweight successor elastically extracted from ERNIE 5.0 (~1/3 total parameters, ~half the active parameters per query) at ~6% of peer pre-training cost; a top Chinese model on LMArena text (API only).

ERNIE 4.0 is an text generation model from Baidu, designed for general AI application workflows.

ERNIE 3.0 is an text generation model from Baidu, designed for general AI application workflows.

ERNIE 3.5 is an text generation model from Baidu, designed for general AI application workflows.

ERNIE 2.0 is an text generation model from Baidu, designed for general AI application workflows.

ERNIE 1.0 is an text generation model from Baidu, designed for general AI application workflows.

Native omni-modal foundation model released January 22, 2026: 2.4T-parameter MoE with an elastic sub-model design jointly modeling text, image, video and audio; 128K context on Qianfan (API only).

Baidu’s newest open-weight family (June 2025): 10 natively multimodal Apache-2.0 variants from 0.3B up to a 424B-total/47B-active MoE, including VL and Thinking variants.

Released September 21, 2026 with a 500K context window at $2/$6 per 1M tokens.

Grok 4 Fast is an text generation model from Xai, designed for general AI application workflows.

Grok Code Fast 1 is an text generation model from Xai, designed for general AI application workflows.

Released August 14, 2026.

Grok 4 is an text generation model from Xai, designed for general AI application workflows.

Grok 4 Heavy is an text generation model from Xai, designed for general AI application workflows.

Grok 3 Mini is an text generation model from Xai, designed for general AI application workflows.

Grok 3 is an text generation model from Xai, designed for general AI application workflows.

Grok 2 is an text generation model from Xai, designed for general AI application workflows.

Grok 2 Mini is an text generation model from Xai, designed for general AI application workflows.

Grok 1.5V is an text generation model from Xai, designed for general AI application workflows.

Grok 1.5 is an text generation model from Xai, designed for general AI application workflows.

New-generation image model (preview) released September 21, 2026 for professional creation: text-to-image and image-to-image with up to 5 reference images, multi-round conversational editing and up to 2K output at roughly $0.024 per image via Tencent Cloud.

1.5B open-source speech foundation model released September 9, 2026: zero-shot TTS and voice cloning, instruct TTS, voice content/acoustic/paralinguistic editing, speech enhancement and separation, with a distilled AuK-Flash variant.

Hunyuan TurboS Latest is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan TurboS 20250716 is an text generation model from Tencent, designed for general AI application workflows.

Next-generation open-source MoE flagship released August 28, 2026: 770B total / 49B activated (256 experts, top-8) with 1M-token context, gated sparse attention with IndexCache and native MTP speculative decoding, tuned for software engineering and research (Apache 2.0).

Hunyuan T1 20250822 is an reasoning model model from Tencent, designed for reasoning, planning, and complex problem solving.

Hunyuan T1 Latest is an reasoning model model from Tencent, designed for reasoning, planning, and complex problem solving.

Hunyuan Functioncall is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Turbo is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan A13B is an reasoning model model from Tencent, designed for reasoning, planning, and complex problem solving.

Tencent Hunyuan Video Generation Effects is an video generation model from Tencent, designed for video generation and motion content production.

Tencent Hunyuan Video Generation is an video generation model from Tencent, designed for video generation and motion content production.

Hunyuan Large Vision is an image understanding model from Tencent, designed for vision understanding and document analysis.

Hunyuan TurboS Vision is an image understanding model from Tencent, designed for vision understanding and document analysis.

Hunyuan T1 Vision is an full modal model from Tencent, designed for general AI application workflows.

Hunyuan Translation Lite is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Standard is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Standard 256K is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Large is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Vision is an image understanding model from Tencent, designed for vision understanding and document analysis.

Hunyuan Standard Vision is an image understanding model from Tencent, designed for vision understanding and document analysis.

Hunyuan Large Longcontext is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Lite Vision is an image understanding model from Tencent, designed for vision understanding and document analysis.

Open-weight MoE LLM fully released July 6, 2026: 295B total / 21B active parameters (+3.8B MTP layer), Tencent’s leading reasoning- and agent-focused model in its class, distributed via Hugging Face and ModelScope.

Hunyuan Code is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Lite is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan Translation is an text generation model from Tencent, designed for general AI application workflows.

Hunyuan TurboS Longtext 128k 20250325 is an text generation model from Tencent, designed for general AI application workflows.

Tencent Hunyuan Image Generation (Multi-turn Dialogue) is an image generation model from Tencent, designed for image generation and creative visual workflows.

Hunyuan Role is an text generation model from Tencent, designed for general AI application workflows.

Text-to-Image Lite is an image generation model from Tencent, designed for image generation and creative visual workflows.

Flagship closed reasoning model released December 5, 2025: MoE with 406B total / 32B active parameters and 256K context, targeting math, science and code (API only).

Lightweight 8.3B open video generation model (November 21, 2025) supporting text-to-video and image-to-video, running on 14GB-VRAM consumer GPUs; still Tencent’s latest open video model.

Hunyuan Embedding is an embedding model from tencent, designed for general AI application workflows.

Tencent Yuanqi is an text generation model from tencent, designed for general AI application workflows.

GLM 4.5 Flash is an text generation model from Chatglm, designed for general AI application workflows.

GLM 4.5 X is an text generation model from Chatglm, designed for general AI application workflows.

GLM 4.5 AirX is an text generation model from Chatglm, designed for general AI application workflows.

GLM 4.5 is an text generation model from Chatglm, designed for general AI application workflows.

GLM 4.5 Air is an text generation model from Chatglm, designed for general AI application workflows.

Efficient tier of the GLM-5.3 generation, released August 26, 2026.

Current Zhipu (Z.ai) flagship released August 14, 2026, with GLM-5.3-Flash alongside.

Cogview 4 is an image generation model from Chatglm, designed for image generation and creative visual workflows.

Cogvideox 3 is an video generation model from Chatglm, designed for video generation and motion content production.

GLM Z1 Air is an reasoning model model from Chatglm, designed for reasoning, planning, and complex problem solving.

GLM Z1 Flash is an reasoning model model from Chatglm, designed for reasoning, planning, and complex problem solving.

GLM Z1 AirX is an reasoning model model from Chatglm, designed for reasoning, planning, and complex problem solving.

Cogvideox Flash is an video generation model from Chatglm, designed for video generation and motion content production.

Cogvideox 2 is an video generation model from Chatglm, designed for video generation and motion content production.

1M lossless context release (~June 2026).

GLM 4 9B is an text generation model from Chatglm, designed for general AI application workflows.

GLM 4 Long is an text generation model from Chatglm, designed for general AI application workflows.

Cogview 3 Plus is an image generation model from Chatglm, designed for image generation and creative visual workflows.

Cogview 3 Flash is an image generation model from Chatglm, designed for image generation and creative visual workflows.

GLM 4 Plus is an text generation model from Chatglm, designed for general AI application workflows.

GLM 4 Assistant is an reasoning model model from Chatglm, designed for reasoning, planning, and complex problem solving.

GLM 4 AllTools is an reasoning model model from Chatglm, designed for reasoning, planning, and complex problem solving.

GLM 4 Flash is an text generation model from Chatglm, designed for general AI application workflows.

GLM 4 FlashX is an text generation model from Chatglm, designed for general AI application workflows.

Cogview 3 is an image generation model from Chatglm, designed for image generation and creative visual workflows.

GLM 4 is an text generation model from Chatglm, designed for general AI application workflows.

GLM 3 Turbo is an text generation model from Chatglm, designed for general AI application workflows.

Flagship released February 11, 2026.

Released December 22, 2025; the free-tier 4.7-Flash followed on January 19, 2026.

Coding and agentic preview rolled out to Kimi Code and Kimi Work on September 11, 2026: overall performance close to Kimi K3 at roughly half the price, with 1M-token context, three thinking-effort levels and image input.

Released July 16, 2026; 2.8T-parameter MoE with open weights following July 27. Agentic coding is a headline use case.

Kimi K2 Turbo Preview is an reasoning model model from Moonshot, designed for reasoning, planning, and complex problem solving.

Kimi K2 0905 Preview is an reasoning model model from Moonshot, designed for reasoning, planning, and complex problem solving.

Kimi Thinking Preview is an reasoning model model from Moonshot, designed for reasoning, planning, and complex problem solving.

Moonshot V1 128k Vision Preview is an image understanding model from Moonshot, designed for vision understanding and document analysis.

Moonshot V1 32k Vision Preview is an image understanding model from Moonshot, designed for vision understanding and document analysis.

Moonshot V1 8k Vision Preview is an image understanding model from Moonshot, designed for vision understanding and document analysis.

Kimi K2 0711 Preview is an reasoning model model from Moonshot, designed for reasoning, planning, and complex problem solving.

Moonshot V1 128k is an text generation model from Moonshot, designed for general AI application workflows.

Moonshot V1 8k is an text generation model from Moonshot, designed for general AI application workflows.

Moonshot V1 32k is an text generation model from Moonshot, designed for general AI application workflows.

Kimi Latest 32k is an text generation model from Moonshot, designed for general AI application workflows.

Kimi Latest 128k is an text generation model from Moonshot, designed for general AI application workflows.

Kimi Latest 8k is an text generation model from Moonshot, designed for general AI application workflows.

Released April 20, 2026.

Kimi K2 is an text generation model from Moonshot, designed for general AI application workflows.

Multimodal coding preview that debuted on MiniMax Code on September 27, 2026 with API public beta the next day: 1M-token context and adjustable thinking effort for everyday coding, frontend and research workloads; free preview ran through October 7.

Omni-modal video model (July 31, 2026): 2K/24fps 15s clips with synchronized audio; open weights with 3.5M+ downloads.

Minimax M1 is an text generation model from Minimax, designed for general AI application workflows.

Minimax Hailuo 02 1080P is an video generation model from Minimax, designed for video generation and motion content production.

Minimax Hailuo 02 512P is an video generation model from Minimax, designed for video generation and motion content production.

Minimax Hailuo 02 768P is an video generation model from Minimax, designed for video generation and motion content production.

Minimax Text 01 is an full modal model from Minimax, designed for general AI application workflows.

Minimax Image 01 is an image generation model from Minimax, designed for image generation and creative visual workflows.

Minimax Image 01 Live is an image generation model from Minimax, designed for image generation and creative visual workflows.

Minimax S2V 01 is an video generation model from Minimax, designed for video generation and motion content production.

Minimax T2V 01 is an video generation model from Minimax, designed for video generation and motion content production.

Minimax I2V 01 is an video generation model from Minimax, designed for video generation and motion content production.

Released February 12, 2026 with open weights; 80.2% SWE-bench Verified at ~$0.30/M input.

Native unified multimodal model open-sourced August 20, 2026 on the NEO-unify architecture (Apache 2.0): native 4K generation, stronger complex-instruction following, text/layout rendering and image editing, with support for 3-4K-character ultra-long instructions.

Flagship delivery-grade natively multimodal agent foundation unveiled July 18, 2026 at WAIC, built for long-horizon tasks such as data analysis, deep research and iterative image editing (API/invite at launch).

Unified vision understanding-and-generation model fully open-sourced July 13, 2026: the 7B Mixture-of-Transformers model reformulates detection, segmentation and depth estimation as multimodal text generation, shipped with a 50M-pair training corpus.

Sensenova V6 Turbo is an full modal model from Sensetime, designed for general AI application workflows.

Sensenova V6 Pro is an full modal model from Sensetime, designed for general AI application workflows.

Sensenova V6 Reasoner is an reasoning model model from Sensetime, designed for reasoning, planning, and complex problem solving.

SenseNova V6.5 Miaohua Base is an image generation model from Sensetime, designed for image generation and creative visual workflows.

SenseNova V6.5 Miaohua Pro is an image generation model from Sensetime, designed for image generation and creative visual workflows.

Natively unified understanding-and-generation multimodal series released open-source in late April 2026 alongside a free Token Plan; iterated as the open U1.5-Lite-Preview toward U1 Pro.

Baichuan M2 32B is an reasoning model model from Baichuan, designed for reasoning, planning, and complex problem solving.

Baichuan2 53B is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan2 Turbo 192k is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan M1 14B is an reasoning model model from Baichuan, designed for reasoning, planning, and complex problem solving.

Baichuan2 Turbo is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan2 13B Chat is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan2 7B Base is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan2 13B Base is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan2 7B Chat is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan 13B is an text generation model from Baichuan, designed for general AI application workflows.

Baichuan 7B is an text generation model from Baichuan, designed for general AI application workflows.

Commercial API model released January 22, 2026 on the M3 base, adding the six-source evidence-based paradigm with factual hallucination reduced to 2.6%.

Open-weight medical-enhanced LLM released January 13, 2026 (~235B): #1 on HealthBench Hard at launch, combining clinical-grade inquiry with low hallucination via domain RL (Apache 2.0).

Yi Lightning is an text generation model from 01-ai, designed for general AI application workflows.

Yi Medium is an text generation model from 01-ai, designed for general AI application workflows.

Yi Large is an text generation model from 01-ai, designed for general AI application workflows.

Yi Large RAG is an text generation model from 01-ai, designed for general AI application workflows.

Yi Large Turbo is an text generation model from 01-ai, designed for general AI application workflows.

Yi 9B is an text generation model from 01-ai, designed for general AI application workflows.

Yi 9B 200K is an text generation model from 01-ai, designed for general AI application workflows.

Yi 6B 200K is an text generation model from 01-ai, designed for general AI application workflows.

Yi 34B 200K is an text generation model from 01-ai, designed for general AI application workflows.

Yi 34B Chat is an text generation model from 01-ai, designed for general AI application workflows.

Yi 6B Chat is an text generation model from 01-ai, designed for general AI application workflows.

Yi 6B is an text generation model from 01-ai, designed for general AI application workflows.

Yi 34B is an text generation model from 01-ai, designed for general AI application workflows.

New flagship MoE released September 7, 2026: 293B total / ~30B active parameters, 200+ languages, major coding and agent upgrades, trained and served fully on domestic compute (API only).

1.7B on-device dense LLM open-sourced (Apache 2.0) on September 1, 2026 alongside the 4B edge model: up to 1M-token native context via hybrid attention, 200+ languages, trained on Huawei Ascend clusters.

Open-weights on-device edge model released September 1, 2026 by Ciyuan Spark; among the first edge models with native 1M-token context (Apache 2.0).

Spark X1 is an text generation model from Iflytek, designed for general AI application workflows.

Spark Tiny is an text generation model from Iflytek, designed for general AI application workflows.

Spark Medical Model Lite is an text generation model from Iflytek, designed for general AI application workflows.

Spark Mini is an text generation model from Iflytek, designed for general AI application workflows.

Spark Lite Patch is an text generation model from Iflytek, designed for general AI application workflows.

Spark Max is an text generation model from Iflytek, designed for general AI application workflows.

Spark Mini Instruct is an text generation model from Iflytek, designed for general AI application workflows.

Deep-reasoning flagship released February 11, 2026, trained entirely on Huawei Ascend compute; ~50% capability uplift over X1.5 with an API-controllable dynamic thinking mode (API only).

Flagship sparse-MoE LLM open-sourced July 31, 2026: 505B total / 18B active parameters with 512K context, released with weights, inference code and a tech report.

Pangu NLP N2 32K 5.0.1.1 is an text generation model from Huawei, designed for general AI application workflows.

Pangu NLP N2 128K 5.0.1.1 is an text generation model from Huawei, designed for general AI application workflows.

Pangu NLP N4 Reasoner 128K 3.0.1.2 is an reasoning model model from Huawei, designed for reasoning, planning, and complex problem solving.

Pangu NLP N2 Reasoner 128K 5.0.0.1 is an reasoning model model from Huawei, designed for reasoning, planning, and complex problem solving.

Pangu NLP N1 Reasoner 128K 3.0.1.1 is an reasoning model model from Huawei, designed for reasoning, planning, and complex problem solving.

Pangu NLP N2 Reasoner 32K 3.0.1.1 is an reasoning model model from Huawei, designed for reasoning, planning, and complex problem solving.

Pangu NLP N4 Reasoner 32K 3.0.1.2 is an reasoning model model from Huawei, designed for reasoning, planning, and complex problem solving.

Pangu AgentExpert N2 0.0.2 is an text generation model from Huawei, designed for general AI application workflows.

Pangu AgentExpert N1 0.0.2 is an text generation model from Huawei, designed for general AI application workflows.

Pangu MM M2 AIGVideo 1.0.0 is an video generation model from Huawei, designed for video generation and motion content production.

Pangu MM M2 Text2Video 1.0.0 is an video generation model from Huawei, designed for video generation and motion content production.

Pangu MM M1 Txt2Img 1.0.0 is an image generation model from Huawei, designed for image generation and creative visual workflows.

Pangu RAG N1 32K 4.3.2 is an text generation model from Huawei, designed for general AI application workflows.

Pangu NLP N4 4K 3.2.36 is an text generation model from Huawei, designed for general AI application workflows.

Pangu NLP N1 128K 3.2.36 is an text generation model from Huawei, designed for general AI application workflows.

Pangu NLP N1 32K 3.2.36 is an text generation model from Huawei, designed for general AI application workflows.

Pangu NLP N2 32K 3.1.35 is an text generation model from Huawei, designed for general AI application workflows.

Pangu NLP N4 32K 2.5.35 is an text generation model from Huawei, designed for general AI application workflows.

92B-total / 6B-active sparse-MoE sibling of Pro with 512K context; first of seven components progressively open-sourced from June 30, 2026.

Starcoder2 3b is an text generation model from Bigcode, designed for general AI application workflows.

Flagship agentic foundation model released September 20, 2026: sparse MoE with 600B total / 27B active parameters, 1M-token context, native text+image+video input and 64K max output at $1.00/$2.70 per 1M tokens; open weights announced for October 15, 2026.

Speech and audio model series launched September 15, 2026: Step-Audio 3 Realtime delivers full-duplex low-latency voice dialogue with adaptive reasoning, alongside ASR, TTS, Music and Gen variants that generate complete soundscapes from a single description.

Step Video TI2V is an video generation model from Stepfun, designed for video generation and motion content production.

Step Video T2V is an video generation model from Stepfun, designed for video generation and motion content production.

Step Video T2V Turbo is an video generation model from Stepfun, designed for video generation and motion content production.

Step 2 is an text generation model from Stepfun, designed for general AI application workflows.

Step 2 Mini is an text generation model from Stepfun, designed for general AI application workflows.

Step 1X is an image generation model from Stepfun, designed for image generation and creative visual workflows.

Open-weight vision-language MoE released May 29, 2026: 198B total / 11B active with ~256K context and an explicit reasoning mode, built for coding agents and search workflows.

~196B sparse-MoE open foundation model released February 2, 2026 (Apache 2.0), positioned as StepFun’s most capable open release for frontier reasoning and agents; base checkpoints also published.