Models

40 labs · 380 models

Newest model · 2 days ago

Scribe v3 Live

ElevenLabs Audio Medium model Closed weights $0.22 per hour

Read ElevenLabs’s announcement ↗

Just announced

The latest models, newest first. Open one for its price, context window and our coverage.

LLMs

All 112

Claude Fable 5.2 Anthropic 3d
Muse Ember Meta 3d
Inkling 2 Small Thinking Machines 7d
Hy4 Tencent 8d
Gemini 4 Carbon Google 15d

Image models

All 81

MiniMax H3-Image MiniMax 2d
Seedream 5.5 ByteDance 3d
GPT-Image-2.6 Flare OpenAI 8d
GPT-Image-2.6 Sunburst OpenAI 8d
P-H3-Image Pruna AI 9d

Video models

All 63

Gemini Omni 1.2 Flash Google 7d
Pika H3 Flex Pika 14d
PixVerse V6 Max PixVerse 21d
Vidu Q4 Preview Shengshu 36d
Kandinsky 6.0 Video Lite Sber 37d

Who’s due next

How far each lab is into its usual gap between releases of a kind — LLMs under All. A full bar means it’s due; overdue only once it’s past its slowest normal gap.

xAI ~40d due now
Google ~29d due in ~14d
Meta ~28d due in ~25d
OpenAI ~49d due in ~32d
Anthropic ~40d due in ~37d
MiniMax ~63d due in ~39d
Xiaomi ~92d due in ~41d
DeepSeek ~74d due in ~48d
Mistral ~104d due in ~67d
Moonshot ~87d due in ~67d

Who announces most

Every announcement in the last 12 months, busiest lab first. Bigger dots are bigger models; rings are open weights. Hover a dot for details.

OpenAI

Google

Alibaba

Microsoft

Mistral

Anthropic

xAI

ElevenLabs

MiniMax

What each lab offers now

The newest model in each of a lab’s lines at every size, and how long ago.

Closed weights

Tier = where the lab puts it in its own lineup

Lab	Big	Medium	Light
ElevenLabs	Scribe v3 Audio 2d SFX v3 Audio 10d Eleven Realtime v1 Max Audio 24d Eleven v4 Audio 45d Eleven Music v2.5 Audio 59d Eleven Multilingual v2 Audio 1178d —	Scribe v3 Live Audio 2d Eleven v3 Conversational Audio 276d —	Eleven Realtime v1 Mini Audio 24d Eleven v4 Turbo Audio 45d Eleven Flash v2.5 Audio 693d —
Anthropic	Claude Mythos 5.2 Cyber 3d Claude Fable 5.2 3d Claude Opus 5.5 51d —	Claude Sonnet 5.5 45d —	Claude Haiku 5.5 36d —
Meta	Muse Wonder 31d Muse Wonder Full 31d Muse Image Image 128d —	Muse Spark 1.3 71d Muse Voice Transcribe Audio 72d —	Muse Ember 3d —
ByteDance	SeedRealtime Audio 100d Seedance 2.5 Video 105d Seed Audio 1.0 Audio 115d Seedream 5.0 Pro Image 127d Seed2.1 Pro 142d Seeduplex Audio 218d —	Seedream 5.5 Image 3d —	Seedream 5.0 Lite Image 272d —
OpenAI	GPT-6 Cyber Cyber 7d GPT-Image-2.6 Sunburst Image 8d GPT-6 Rosalind Biology 13d GPT-6 Astra 70d GPT-Live-1 Audio 127d GPT-Realtime-2.1 Audio 129d GPT-Audio-1.5 Audio 262d GPT-5.3-Codex Coding 280d —	GPT-Image-2.6 Flare Image 8d GPT-6.1 Terra 17d GPT-Transcribe Audio 107d GPT-Realtime-Translate Audio 189d GPT-Rosalind Biology 210d GPT-5.1-Codex Coding 364d —	GPT-6 Luna 51d GPT-Live-Transcribe Audio 107d GPT-Live-1 mini Audio 127d GPT-Realtime-2.1 mini Audio 129d GPT-5.1-Codex-Mini Coding 364d gpt-image-1-mini Image 402d —
Google	Gemini 4 Carbon 15d Gemini 4 Argon 43d Gemini 3.8 Live Extended Thinking Audio 58d Lyria 3.5 Audio 106d Gemini 3.1 Pro 266d Nano Banana Pro Image 357d Veo 3.1 Video 393d —	Gemini Omni 1.2 Flash Video 7d Gemini Nano Banana 2.1 Image 37d Gemini 3.8 Flash TTS Audio 51d Gemini 3.8 Live Audio 58d Gemini 3.8 Flash Cyber Cyber 71d Gemini 3.8 Flash 71d Gemini 3.5 Transcribe Live Audio 78d Lyria 3 Audio 267d Veo 3.1 Fast Video 393d Chirp 3: Transcription Audio 580d —	Gemini 3.8 Flash-Lite TTS Audio 51d Gemini 3.5 Flash-Lite 114d Nano Banana 2 Lite Image 135d Veo 3.1 Lite Video 226d —
Thinking Machines	—	—	Inkling 2 Small 594B 7d —
Microsoft	MAI Image 2.7 Image 16d MAI-Voice-2.1 Audio 42d —	MAI-Transcribe-2-Streaming Audio 42d MAI-Thinking-1 ~1T total, 35B active 92d MAI-Image-2.5 Image 163d —	MAI-Cyber-1.1-Flash Cyber 16d MAI-Voice-2.1-Flash Audio 42d MAI-Image-2.6-Flash Image 69d MAI-Code-1-Flash Coding 5B active 163d —
PixVerse	PixVerse V6 Max Video 21d PixVerse R2 Video 51d —	—	—
Mistral	Voxtral Small Transcribe 3 Audio 23d —	Codestral 25.08 Coding 470d —	Voxtral TTS 3 Audio 23d Mistral Small Creative Companion 336d —
Krea	Krea 3 Image 28d —	Krea 2 Medium Image 184d —	—
xAI	Grok Voice 3.0 Audio 29d Grok 4.7 52d Grok Imagine Image 2.0 Image 97d Grok Imagine Video 1.5 Video 162d —	Grok Voice Transcribe 2.0 Audio 56d Grok Imagine Image Image 288d —	Grok Imagine Video 1.5 Fast Video 149d Grok Build 0.1 Coding 177d Grok 4.1 Fast 358d —
Shengshu	Vidu Q4 Preview Video 36d —	Vidu S1 Video 132d —	Vidu Q3 Turbo Video 274d —
Black Forest Labs	FLUX 3 Image Image 42d FLUX 3 Video Video 112d —	FLUX.2 [flex] Image 352d —	FLUX Video Upscale Video 84d —
Ideogram	Ideogram 4.5 Image 43d —	—	—
Kling	Kling 4.0 Video 45d Kling VIDEO 3.0 Omni Video 280d Kling Image 3.0 Omni Image 280d —	Kling Image 3.0 Image 280d —	Kling 4.0 Flash Video 45d —
Recraft	Recraft V4.1 Image 182d —	Recraft V4 Styles Image 78d —	Recraft V4.1 Flash Image 50d —
Alibaba	Wan3.0 Video 98d Qwen-Image-3.0 Image 119d —	Qwen3.8-LiveTranslate Audio 55d Qwen-Image-2.0 Image 278d Qwen3-TTS-VD-Flash Audio 324d —	Wan3.0-Video-Prime Video 84d —
Runway	Aleph 2.0 Video 175d Gen-4.5 Video 346d Frames Image 717d —	—	Enhance Frame Rate Video 56d Gen-4 Image Turbo Image 450d Gen-4 Turbo Video 584d —
Midjourney	Edit Model for V8 Image 77d Niji V7 Image 307d —	—	—
Pika	—	—	Pika Soundtrack Audio 90d Pika SFX Audio 90d Pika Music Audio 90d Pika Speech Audio 3B 90d PikaStream1.0 Video 9B 224d —
China Mobile	—	JT-4.1 Flash 236B 118d —	—
Reve	Reve 2.1 Image 126d —	—	—
Aorizon	HappyHorse-1.1 Video 149d —	—	—
Luma AI	Ray3.2 Video 156d Uni-1.1 Max Image 191d —	Uni-1.1 Image 191d —	—
StepFun	—	—	Step Image Edit 2 Image 197d —
Sber	Kandinsky 6.0 Image Pro Image 6B 198d —	—	—
Xiaomi	Xiaomi MiMo-V2-Pro 1T 239d —	MiMo-V2.5-TTS Audio 203d Xiaomi MiMo-V2-Omni 239d —	—
MiniMax	MiniMax Music 2.6 Audio 216d MiniMax Speech 2.8 HD Audio 293d MiniMax Hailuo 2.3 Video 380d —	MiniMax-M2-her Roleplay 289d —	MiniMax Speech 2.8 Turbo Audio 293d MiniMax Hailuo 2.3 Fast Video 380d —
Leonardo AI	Lucid Origin Image 464d —	—	—

Open weights

Tier by size · Big 200B+ · Medium 20–200B · Light under 20B

Lab	Big	Medium	Light
MiniMax	MiniMax H3-Image Image 2d MiniMax H3 Video 104d MiniMax Music 3.0 Audio 119d MiniMax M3 428B 165d MiniMax M2.7 239d —	—	MiniMax M3.1 Flash 24d —
Tencent	Hy4 8d HunyuanImage-3.0-Instruct Image 80B (13B active) 290d HunyuanVideo 1.5 Video 8.3B 357d —	—	HunyuanImage-3.0-Instruct-Distil Image 80B (13B active) 290d —
Pruna AI	—	P-H3-Image Image 20B 9d —	—
Pika	—	—	Pika H3 Flex Video 14d —
NVIDIA	Nemotron 3.5 Transcribe 30B-A4B Audio 34B (4.3B active) 15d Cosmos 3 Super Video 64B 165d Cosmos3-Super-Text2Image Image 64B 165d —	NemotronLabs VoiceChat 11B Audio 11B 101d Cosmos 3 Nano Video 16B 165d —	Cosmos 3 Edge Video 4B 115d Nemotron 3.5 ASR Streaming 0.6B Audio 600M 161d Parakeet Unified EN 0.6B Audio 600M 219d Canary-1B-v2 Audio 1B 455d —
Moonshot	Kimi K3.1 Pro 2.8T 20d Kimi K2.7 Code Coding 1T-A32B 154d —	—	—
DeepSeek	DeepSeek-V4.1-Flash-1017 552B 26d DeepSeek-V4-Pro-0813 1.6T 91d DeepSeek-Math-V2 Math 685B 350d —	—	—
Black Forest Labs	—	FLUX.2 [dev] Image 32B 352d —	FLUX 3 Image Klein Image 5.6B 28d —
Meta	Muse Spark Open 816B 28d —	Muse Glimmer 30B 94d Omnilingual ASR Audio 300M–7B 367d —	—
Microsoft	—	VibeVoice-ASR-Streaming Audio 1.5B / 7B 70d —	MAI-Code-1.1-Flash Coding 36d VibeVoice-Realtime-0.5B Audio 0.5B 344d —
Sber	Kandinsky 6.0 Video Pro Video 29B 37d —	—	Kandinsky 6.0 Video Lite Video 3B 37d Kandinsky 5.0 Image Lite Image 6B 349d —
Mistral	Mistral Large 4 1.05T 37d Devstral 2 Coding 123B 338d —	Leanstral 1.5 Math 119B-A6B 133d Mistral Medium 3.5 128B 174d Mistral Small 4 119B 241d Voxtral Small Audio 24B 485d —	Voxtral TTS Audio 4B 234d Devstral Small 2 Coding 24B 338d Ministral 3 14B 14B 345d —
Xiaomi	MiMo-V2.6-Pro 51d —	MiMo-V2.5-ASR Audio 203d —	—
Alibaba	Qwen3.8-Max 2.4T 102d —	Qwen3.8-Flash-Next 125B 78d Qwen3-Coder-Next Coding 80B-A3B 282d Z-Image Image 6B 289d Qwen-Image-2512 Image 317d —	Qwen-Image-2.1 Image 7B 53d Qwen3-ASR Audio 0.6B / 1.7B 287d Qwen3-TTS Audio 0.6B / 1.7B 294d Z-Image-Turbo Image 6B 351d —
Z.ai	GLM-5.3 744B 90d GLM-Image Image 9B + 7B 302d —	—	GLM-TTS Audio 336d GLM-ASR-Nano-2512 Audio 1.5B 338d —
Krea	—	Krea 2 Raw Image 12B 142d —	Krea 2 Turbo Image 12B 162d —
Ideogram	Ideogram 4.0 Image 9.3B 162d —	—	—
HiDream	HiDream-O1-Image Image 8B 188d —	—	HiDream-O1-Image-Dev Image 8B 188d —
LG AI Research	—	EXAONE 4.5 33B 217d —	—
Google	—	Gemma 4 31B 31B 224d —	Gemma 4 E4B 4B 224d —
Skywork	SkyReels V3 Video 287d —	—	—
StepFun	—	Step1X-Edit-v1p2 Image 351d —	—
OpenAI	—	gpt-oss-20b 21B 464d gpt-oss-120b 117B 464d —	—
Genmo	—	—	Mochi 1 Video 10B 751d —

Days count from the day a lab announced the model — previews count, leaks and teasers don’t. LLMs, plus the labs’ cyber, specialist (coding, math, companion), image, video and audio models. “Usual gap” is the median time between a lab’s releases of a kind over the last 12 months — launches within two weeks of each other count as one release, and a lab needs four releases to have one. It’s only overdue once it’s past its slowest normal gap.

Scribe v3 Live

Claude Fable 5.2

Claude Mythos 5.2

GPT-6 Rosalind

MiniMax H3-Image

Gemini Omni 1.2 Flash

Scribe v3 Live