RADAR ARCHIVE
Week 18, 2026
9 tools added to the directory, Apr 27 – May 3, 2026.
Added to the directory (9)
Tested and written up.
Alibaba flagship video with joint audio and multilingual lip-sync
Alibaba's HappyHorse 1
Why: It addresses the hardest user complaint about AI video, convincing sound and lip-sync with motion, not only pixels. Strong fit when you need dialogue-forward clips or localized performances without a full audio post stack.
Paid
Best for Audio+Video
Visit
One multimodal model for text, vision, audio, and video reasoning
Nemotron 3 Nano Omni is NVIDIA's compact-but-capable multimodal stack for agentic workflows: one family of endpoints that accept text, images, audio, or video (depending on route) and return text answ...
Why: If your product roadmap says 'agents that see and hear the world,' Omni is built for that integration story, fewer moving parts than bolting Whisper + CLIP + LLM together by hand.
Paid
Best for Agents
Visit
Multi-image to production-grade 3D on next-gen Meshy
Meshy 6 continues Meshy's focus on fast, usable 3D assets with emphasis on multi-image conditioning: feed several views or references so the model better infers shape, materials, and proportions for g...
Why: Teams outgrew 'cool sculpt from one photo' and need consistent assets from multiple references, Meshy 6 is explicitly positioned for that workflow.
Freemium
Best for Multi-Ref 3D
Visit
Real-time virtual try-on in video
Lucy 2
Why: Most directories list generic video models; few spell out 'commerce motion' workflows, VTON fills that gap for teams selling apparel and accessories.
Paid
Best for Fashion Video
Visit
Recursive self-improvement language model for real-world engineering
MiniMax M2
Why: MiniMax M2.7 is the current production language model below M3 and is explicitly listed as beginning recursive self-improvement, making it a notable addition to the family.
Freemium
Best for Engineering Tasks
Visit
Same M2.7 performance with significantly faster inference
MiniMax M2
Why: The Highspeed variant is a current, actively promoted option for developers who need M2.7 capability with lower latency.
Freemium
Best for Low-Latency Coding
Visit
Unified open-source small model for chat, reasoning, vision, and coding
A 119B-parameter MoE model with 6B active parameters and a 256K context window, released under Apache 2
Why: Small 4 packs flagship-class reasoning, vision, and coding into a single open-source model that is efficient enough for high-throughput and local deployments.
Freemium
Best for Efficient Open Multimodal
Visit
Creative upscaling that adds realism to AI-generated images
Reinterprets and enhances AI-generated images and digital art up to 8x, adding texture, lighting, and realism while preserving the original composition
Why: A dedicated creative upscaler for AI-generated imagery, bridging the gap between raw AI output and production-ready assets.
Paid
Best for AI Image Polish
Visit
Topaz image enhancement on iPhone
Brings Topaz Photo AI enhancement capabilities to iPhone, allowing mobile photographers to upscale, sharpen, denoise, and enhance images directly on their device
Why: Extends Topaz's photo enhancement models to iPhone, giving mobile creators access to desktop-quality AI polish.
Freemium
Best for Mobile Photo Enhancement
Visit