Arcee Ships Open Reasoning Model at 96% Less Cost
Arcee AI releases Trinity-Large-Thinking under Apache 2.0, ranking #2 on PinchBench behind Claude Opus 4.6 at just $0.90 per million output tokens.
Falcon Perception: 0.6B Vision Model Beats Meta SAM 3
Technology Innovation Institute releases Falcon Perception, a compact 0.6B parameter vision model that beats Meta SAM 3 across six key benchmarks while running on consumer hardware.
PrismML Open-Sources Bonsai 8B: A True 1-Bit LLM
PrismML emerged from stealth and open-sourced Bonsai 8B, a true 1-bit large language model that fits 8.2 billion parameters into just 1.15 GB of memory, roughly 14x more compressed than comparable 16-bit models.
H Company Holo3 Sets New SOTA for GUI Agents
H Company released Holo3, a vision-language model optimized for GUI agents that sets new state-of-the-art on the OSWorld-Verified benchmark, beating proprietary models with only 3B active parameters.
Skywork Matrix-Game-3.0: Open-Source Real-Time Video AI
Skywork AI releases Matrix-Game-3.0, an open-source model that generates real-time interactive video at 720p and 40fps.
NVIDIA Kimodo: Open-Source Motion AI for 3D
NVIDIA has open-sourced Kimodo, the largest controllable motion diffusion model ever trained. Trained on over 700 hours of professional motion capture data, it generates 3D human and robot motions from text prompts.
Voxtral TTS: Open Weights Challenge ElevenLabs
Mistral Voxtral TTS is a 4B parameter open-weights model that matches ElevenLabs quality in human evaluations, with 3-second voice cloning across 9 languages and self-hosting support.
MagiHuman 15B Ships Open-Source Human Video AI
Researchers released daVinci-MagiHuman, a 15-billion-parameter open-source model that generates realistic human video with synchronized speech in two seconds.
Open-Source vs Closed AI: The Creator's Guide for 2026
The performance gap between open-source and closed AI models has collapsed from 17.5 percentage points to effectively zero on knowledge benchmarks by early 2026. For creators, that shift changes everything.
OpenAI Parameter Golf Offers $1M for AI Compression
OpenAI has launched Parameter Golf, an open research competition offering $1M in computing credits to whoever can compress AI models into just 16MB with only 10 minutes of training time.
CorridorKey Ships Open-Source AI Green Screen Tool
Corridor Digital co-founder Niko Pueringer releases CorridorKey, an AI chroma keyer that handles hair, motion blur, and transparency at VFX-production quality.
CorridorKey: When a VFX Artist Builds the Tool
Corridor Digital co-founder Niko Pueringer released CorridorKey, an open-source AI chroma keyer that handles motion blur, hair, and translucent edges at professional quality.
SD3.5-Flash Brings 4-Step Image Generation to Phones
SD3.5-Flash generates images in just 4 steps instead of the usual 30 to 50, making on-device AI image generation practical on smartphones and laptops with under 8GB of RAM.
Google Colab MCP Server Connects AI Agents to GPUs
Google has released an open-source MCP server for Google Colab, letting AI agents like Claude Code and Gemini CLI create notebooks, execute Python code, and access cloud GPUs directly.
Unsloth Studio Brings No-Code LLM Training to Desktop
Unsloth AI released Studio, a free local application that provides a visual interface for fine-tuning large language models while using 70% less VRAM than standard training methods.
Holotron-12B Ships Open-Source Computer Use at 2x Speed
H Company released Holotron-12B on March 17, an open-source computer use agent model that delivers more than 2x the throughput of its predecessor on a single H100 GPU.
Mamba-3: Can SSMs Finally Challenge Transformers?
Mamba-3 matches transformer quality at 1.5B parameters while running inference 7x faster. Three technical innovations close the SSM capability gap.
Mamba-3 Open-Source SSM Rivals Transformers at 7x Speed
Researchers from CMU, Princeton, Together AI, and Cartesia AI published Mamba-3, a state-space model architecture that matches transformer performance while running inference 7x faster.
Leanstral Proves AI Code Is Correct at 93% Less Cost
Mistral released Leanstral, the first open-source AI agent built for Lean 4 formal verification, proving AI-generated code meets its specifications at 93% lower cost.
NVIDIA Nemotron Coalition Unites 8 Labs for Open AI
NVIDIA launched the Nemotron Coalition at GTC, uniting Black Forest Labs, Cursor, Mistral AI, Perplexity, and four other labs to co-develop open frontier AI models.
Mistral Small 4 Ships 119B Open Multimodal Model
Mistral AI released Mistral Small 4, a 119B mixture-of-experts model that unifies reasoning, multimodal vision, and coding under a single Apache 2.0 license.
Open-Source Creative AI Catches Up in March 2026
Six open-source releases in two weeks just made a full creative pipeline possible without a single subscription Between March 1 and March 15, 2026, something happened that has never happened before in creative AI: open-source models matched or exceeded commercial tools across every major creative...
NVIDIA Nemotron 3 Super: Open-Source 120B Model
NVIDIA released Nemotron 3 Super on March 11, 2026, an open-source 120B-parameter hybrid Mamba-Transformer model that delivers 5x higher throughput than its predecessor for agentic AI workloads
Hume TADA: Open-Source TTS With Zero Hallucinations
Hume AI has released TADA (Text-Acoustic Dual-Alignment), an open-source text-to-speech model that eliminates a persistent problem in AI speech synthesis: hallucinated words