AI Model Release Tracker

Intelligent routing to optimal models for 3x cost savings on inference

Gemini 3.5-powered speech-to-text with contextual accuracy for technical/specialized content

7B quantized model for offline laptop deployment, competes with Meta on-device push

Open-weight video/world model, 10s clips from images in 6.8s

Lightweight models with 1B+ downloads and thriving 100K+ derivative ecosystem

SpaceX's reasoning model, competitive with frontier LLMs

30B on-device AI agent, runs natively on consumer hardware

Fast reasoning model with built-in text watermarking for compliance

Anthropic's most advanced reasoning model for complex multi-step tasks

118B MoE model beats rivals 10x its size on coding benchmarks

30-second 4K video with native audio and up to 50 reference inputs

Anthropic's frontier model, currently first on the Artificial Analysis Intelligence Index

Node workflows without running your own GPU

One canvas, many models, wired together

Open-source node canvas built around the edit, not the prompt

Anonymous 1M-context reasoning model available free through OpenRouter

Google's coding workhorse, three weeks after 3.6 Flash

Z.ai's post-trained coding and agentic model on the GLM-5.2 base

Meta's closed-weight agentic model, and its first paid model API

NVIDIA's 550B open-weights reasoning model, built for inference speed

Multimodal model generating image, video and audio from one set of weights

753B open-weight MoE coding model with a 1M-token context, MIT licensed

Thinking Machines' 975B Apache-2.0 model that takes text, images and audio natively

Anthropic's Mythos-class creative model

OpenAI's top-tier model for complex professional work

Moonshot AI's 2.8-trillion-parameter open-weight flagship

xAI's flagship coding model, trained in partnership with Cursor

Omni-modal video with native stereo audio, at 2K

Open-source, model-agnostic terminal coding agent

Moonshot's specialized coding model

Clean, controllable game-ready topology in ~10 seconds

Conversational AI agent for end-to-end 3D creation

Controllable cinematic video model with multi-keyframe direction and motion transfer

OpenAI's balanced GPT-5.6 model for intelligence and cost

OpenAI's cost-optimized GPT-5.6 model for high-volume workloads

Microsoft Research's open image-to-3D model

Microsoft's unified AI model family from Build 2026

Open physical-AI omnimodel for robotics and AV

Hosted Alibaba TTS across 16 languages, in a fast tier and a fidelity tier

Google's faster, sharper agentic-coding upgrade to 3.5 Flash

Alibaba's 2.4-trillion-parameter flagship, currently in preview

Anthropic's cheaper, near-Opus everyday model

Multimodal reasoning model that generates brand-consistent images and edits

Faster inference variant of Kimi's coding specialist

Limited-availability Mythos-class model without Fable 5 safety classifiers

Compact 30B open-weight model with configurable reasoning for agents

120B open-weight hybrid MoE for efficient multi-agent reasoning

Multimodal 4B safety model for text and image moderation

Open foundation model for humanoid robot reasoning and control

Optimized GLM-5 variant for fast sequential task execution

Multimodal coding and visual-reasoning agent model

Ultra-realistic multilingual text-to-speech with sound tags

Fast multilingual text-to-speech with natural flow

Music generation with humanized vocals and elevated sound

Efficient 8B physical-AI omni-model for workstations

4B physical-AI omni-model for real-time edge robotics

OpenAI's latest image generation model

The node graph the rest of the field is measured against

Anthropic's powerful enterprise model from May 2026

OpenAI's fast default ChatGPT model from May 2026

Cursor's agentic coding model for multi-file software engineering

Google's fast, capable multimodal model from I/O 2026

xAI's agentic coding CLI for autonomous software engineering

Google's unified multimodal generation model

In-context video editing model and Edit Studio

Google's personal AI agent for proactive assistance

Mistral's unified work and coding agent

DeepSeek's open-weight model with permanent pricing

Black Forest Labs' top-tier image generation model

MiniMax's 1M-context agentic frontier model

StepFun's 198B MoE vision-language model

Design and brand image generation with vector support

AI music generation with professional controls

Google's fast text-to-image model via Fal

AI-powered video dubbing in multiple languages

Mistral's mid-tier workhorse for reasoning, coding, and instruction

AI music generation with stems and inpainting

Recraft's most advanced image model with photorealistic, vector, and utility variants

Cloud AI video enhancement up to 4K

Alibaba flagship video with joint audio and multilingual lip-sync

One multimodal model for text, vision, audio, and video reasoning

Multi-image to production-grade 3D on next-gen Meshy

Real-time virtual try-on in video

Recursive self-improvement language model for real-world engineering

Same M2.7 performance with significantly faster inference

Unified open-source small model for chat, reasoning, vision, and coding

Fast text-to-3D and image-to-3D generation

Hyper3D's image-to-3D generation model

Creative upscaling that adds realism to AI-generated images

Topaz image enhancement on iPhone

High-volume DeepSeek inference with a 1M-token context window

Tencent's latest open-source MoE flagship with tool use

Moonshot's open-weight multimodal successor with long-context coding stability

Frontier Opus model with higher-resolution vision and xhigh effort

Mistral's flagship open-weight multimodal frontier model

Mistral's edge family of small, dense open-source models

MIT-licensed MoE flagship for 8-hour autonomous coding sessions

Free universal GLM model with a 200K context window

Free vision model for image understanding and document snapshots

xAI's long-context flagship with a 1M-token window

ByteDance's Next-Gen Video Model with Native Audio-Video Joint Generation

Google's AI Research Assistant: The Ultimate Study Tool

The frontier of cinematic video synthesis

Google's state-of-the-art video generation model

xAI's real-time AI assistant

The Efficiency Revolution: Frontier Intelligence at 1/100th the Cost

Text-to-music & vocals with fast iteration

Meta's open-source large language model

European open-source and commercial LLM

High-quality TTS and voice tools

Enterprise-focused LLM platform

Alibaba's multilingual open-source LLM

Microsoft's efficient small language models

Google's open-source lightweight LLM

Databricks' high-performance open-source LLM

Voice generation and cloning tools

OpenAI's state-of-the-art video model with audio

Top-tier image-to-video with native audio generation

Text/image-to-video generation (availability varies)

Text/image-to-video creation suite with editing tools

Text/image-to-video with Pikaffects (squish, melt, explode)

Avatar and talking-head video generation

Fast video generation from Luma Dream Machine

AI avatar video creation for teams

Fast 1080p image-to-video from MiniMax

The ceiling of enterprise autonomy with 1M context

High-end image generation with strong aesthetics

Audio-driven human animation from ByteDance

Talking avatar videos from images and scripts

Open-source image-to-video with LoRA support

Tencent's high-quality open video model

Text-to-image with strong typography (varies by model)

Image generation with workflows and models

Latest Wan model for text-to-video generation

Stylized image/video animation for creators

Tencent's latest text-to-video model

Generative image tools inside Adobe ecosystem

Fast text-to-video with audio support

Tencent's high-quality 3D generation engine

Creative image workflows (and some video features)

Text/image-to-video with effects, transitions & swaps

Shengshu's advanced image-to-video with better control

Character motion and meme-style video creation

OpenAI's high-fidelity image generation

744B-parameter open-weight MoE flagship for agentic planning and execution

80B parameter open-weight coding powerhouse

The frontier model for complex reasoning and software architecture

The industry standard for coding and nuanced instruction following

The industry standard for cinematic AI video generation

Native multimodal intelligence with a 10M context window

The conversational search engine that replaced traditional search

AI search engine for peer-reviewed scientific research

Instant high-quality 3D modeling from text and images

High-fidelity 3D asset generation from Luma Labs

High-speed, high-realism video generation

The new gold standard for prompt adherence and text rendering

The creative suite for physics-defying video effects

Production-ready 3D assets in under 60 seconds

Meta's Segment Anything 3D for high-fidelity reconstruction

Generate and refine 3D assets from text or images

Fine-tuned control with adjustable inference

xAI's 2M-context beta model with multi-agent capabilities

Design-forward image generation (logos, vectors, assets)

Context-aware image generation and editing

Open-source image generation with flexibility

AI upscaling and enhancement for images

Latest Wan for image variations and editing

FLUX image model family (provider site)

High-fidelity object removal from images

Microsoft's advanced 3D generation from text or images

Object removal from video with high fidelity

2D-to-3D conversion for game assets

Relight and recamera videos

Advanced video editing and effects

Advanced AI music generation with high-quality compositions

High-quality music and sound effects generation

3D capture + creative tools (incl. 3D/Video features)

Multilingual text-to-speech with natural voice synthesis

Open image generation ecosystem (model + tools)

Google's latest music generation model

Design suite with built-in AI generation features

CD-quality music with superior vocals

Audio/video editing with AI features

Advanced sound effects generation

Fast Flux variant for rapid image generation

Google's high-quality text-to-image model

Vector art and brand-style image generation

Exceptional typography and text rendering

Image enhancement (denoise/sharpen/upscale)

Development Flux for advanced control

3D design tool (with AI features depending on product)

Quick text rendering for marketing graphics

Multilingual text rendering and photorealism

7B multimodal model for text and images

Photorealistic Flux with LoRA fine-tuning

Customizable Flux with LoRA fine-tuning

Multilingual text-to-speech with streaming

OpenAI's conditional 3D model generation

OpenAI's fast point cloud generation

Text-to-3D via NeRF with score distillation

NVIDIA's high-quality 3D mesh generation

Professional video upscaling and enhancement

AI-powered video editing with enhancement features

View-consistent image-to-3D generation

Fast single-image 3D generation

The 'Next DeepSeek' Movement: o1-Level Reasoning at 1/100th the Cost

The Open Vision-Reasoner: SOTA Multimodal Performance

Meta's Open Multimodal Standard

The Open Vision Frontier: 124B Multimodal Power

The Open-Source Vision Giant: 78B Multimodal Leader

The Workflow Canvas: Figma for Generative AI

Ultra-fast photorealistic image generation with bilingual text rendering

Open-source 20B model with commercial-grade text rendering and advanced image editing

The Open Image Standard: The Midjourney Killer

Moonshot's open-weight multimodal generalist with agent swarms

Tencent's fast, cost-efficient flagship Hunyuan model

Open-source MoE LLM with strong Chinese NLP and multimodal capabilities

Advanced multilingual LLM with enhanced reasoning and long-context support

Open-source text-to-3D motion model with 200+ motion categories and production-ready exports

Autonomous AI agent for complex multi-step workflows and research automation

Cinematic audio-video joint generation with lip-sync and dialect support

The 128K-context MoE flagship that introduced sparse attention

Strong general-reasoning model with interleaved thinking

Vision-language model for visual reasoning and UI replication

Tripo's latest high-fidelity 3D generation model

First Claude model with the effort parameter and context compaction

Tencent's efficient small-scale MoE instruct model

Fast, lightweight FIBO pipeline designed for speed, efficiency, and on-prem deployment

Cursor's first-generation agentic coding model

xAI's high-volume, 2M-context workhorse model

Fast, low-cost Claude model with extended thinking and Computer Use

Document parsing model for PDF and image OCR

Balanced Sonnet model with major coding and agentic improvements

Tencent's 80B-parameter open-source MoE image generator

The deep-thinking variant of Hunyuan 2.0

Mid-range coding and tool-calling model with 200K context

Mid-generation upgrade between Tripo 3 and 4

xAI's fast, cheap coding specialist model

Tencent's open-source immersive 3D world generator

High-accuracy background removal model trained on a licensed, professionally labeled dataset

Cost-efficient reasoning, coding, and agent model

Mistral's open-weight speech understanding and TTS models

Open-source JSON-native text-to-image model built for controllable, enterprise-safe generation

Tencent's general-purpose instruction-tuned Hunyuan 2.0 model

Tencent's latest open-source high-fidelity 3D asset generator

Fast, inference-efficient 1080p video with native multi-shot storytelling

Fast local FLUX.2 generation for personal hardware

Open-weight FLUX.2 for research and commercial use

Generate custom synthetic voices from text descriptions

Layout-aware document parsing that goes beyond OCR

Pika's refined model with stronger realism and camera control

Browser-based AI image enhancement workflows

Alibaba's open-source MoE flagship with thinking modes

Meta's open-weight flagship with native multimodal reasoning

Long-context, efficient open multimodal model for edge and single-GPU use

Premium tier of Kling 3.0 with best quality

Runway's next-generation model for consistent characters and camera

Google's high-performance reasoning model with advanced coding

Tencent's Mamba-powered deep-thinking reasoning model

Google's open multimodal model for research and developers

Earlier open-source Wan video generation model

Kling's image generation model with style control

Autonomous research agent that performs multi-source deep dives

Alibaba's closed-API flagship before Qwen 3

Chain-of-thought reasoning model for multi-step logical analysis

The open-weight reasoning model that sparked the efficiency revolution

Improved physics and expressive movement in Kling video

Google's high-quality 1080p video generation model

Google's low-latency agentic model with native tool use

Efficient 70B open model matching 405B quality

Ultra-low-latency text-to-speech for real-time voice agents

NVIDIA-aligned 253B Llama 3.1 for helpfulness and instruction following

NVIDIA-aligned 49B Llama 3.1 for balanced performance

NVIDIA-aligned 8B Llama 3.1 for efficient inference

Compact open-source image-to-3D model from Microsoft

High-quality open-source image-to-3D from Microsoft

Alibaba's open coding-specialist model

Tencent's 389B-parameter open-source MoE language model

Advanced search model with deeper reasoning and richer citations

Performance-driven character animation from video

Stability AI's largest 3.5 model with best quality

Ultra-realistic FLUX.1 update with faster generation

Advanced inpainting and outpainting FLUX model

Canny-edge-guided image generation and editing

Depth-map-guided image generation and editing

Kling's standard model for cinematic video

Pika's upgrade with improved motion and effects

Image generation model with strong style control

Fast, low-cost generation for rapid creative exploration

Open mathematical reasoning specialist

Realistic images, flexible styles, and reliable typography in one prompt

Faster, cheaper Gen-3 Alpha for rapid video iteration

The first frontier-scale open-weight language model

Stability AI's first multimodal-diffusion Transformer image model

Efficient SD3 variant for consumer hardware

Kling's first widely available video generation model

Real-time multilingual voice conversion that preserves emotion and content

AI-powered image upscaling up to 8x with detail recovery

Earlier generation of Tripo's text- and image-to-3D pipeline

Fast, cost-efficient multimodal model with a 1M context window

Tencent's open-source bilingual text-to-image diffusion transformer

Second-generation designer-first image generation model

Fast open-source image-to-3D from Stability AI and Tripo

Google's long-context multimodal flagship with up to 2M tokens

Alibaba's strongest vision-language model

Lightweight, real-time search model for cited answers at low cost

Fast text- and image-to-3D for concept exploration

Google's photorealistic text-to-image model with text rendering

Pika's first public video generation model

Fast one-step SDXL for real-time generation

AI assistant embedded across Word, Excel, PowerPoint, Outlook, and Teams

Microsoft's everyday AI assistant across web, PC, and mobile

Emotionally-aware multilingual text-to-speech across 29 languages

High-resolution open-source image generation

Runway's first generation of text- and image-to-video

The foundational open-source text-to-image model

Free AI design and image generation app powered by DALL-E