Articles Tagged "Multimodal"

Bonsai 27B

Bonsai 27B compresses Alibaba's Qwen3.6-27B into 1-bit and ternary weights, shrinking a 54GB model to as little as 3.9GB so it runs on an iPhone.

Moonshot AI's Kimi K3 is a 2.8 trillion parameter MoE model that tops LMArena's Frontend Code Arena and nears Claude Fable 5 on intelligence benchmarks, but at roughly triple Kimi K2.6's price and a higher hallucination rate.

Elorian's $300M Joins AI's No-Product Funding Club

Ex-Google DeepMind researcher Andrew Dai raised $55M at a $300M valuation for Elorian, a visual AI lab with no product yet, joining a growing list of frontier labs priced on pedigree alone.

Inkling

Thinking Machines Lab's first open-weight model - a 975B-parameter MoE with native text, image, and audio reasoning, released under Apache 2.0 and tuned for customization on the Tinker platform.

Gemini 3 Pro

Google DeepMind's Gemini 3 Pro debuted at 1501 Elo on LMArena with 91.9% on GPQA Diamond and a 1M-token context window, before Google retired it for Gemini 3.1 Pro.

Gemini 2.5 Flash

Google's hybrid reasoning workhorse pairs a 1M-token context window with $0.30/$2.50 per million token pricing and a toggleable 0-24,576 token thinking budget, now heading toward an October 2026 shutdown.

Seedream 5.0 Pro

ByteDance's July 2026 flagship image generation model with native 2K output, editable layer separation, multilingual text in 14 languages, and precision region editing.

ByteDance Seedream 5.0 Pro - 2x Faster Than GPT-Image 2

ByteDance's Seedream 5.0 Pro reaches developer APIs with 2K output, 14-language text rendering, and layer-based editing at roughly one-fifth the cost of GPT-Image 2.

GPT-Live-1

OpenAI's full-duplex voice model that listens and speaks simultaneously, replacing Advanced Voice Mode in ChatGPT with three reasoning tiers backed by GPT-5.5.

Gemini Omni Flash

Google DeepMind's multimodal video generation model that creates 10-second clips with native audio from text, images, or video inputs - and lets you refine results through conversation.

Holo3-35B-A3B

H Company's open-weight sparse MoE vision-language model purpose-built for desktop computer use, scoring 82.6% on OSWorld-Verified with only 3B active parameters.

Gemini 3.5 Pro

Google DeepMind's upcoming flagship model with a 2M-token context window and Deep Think reasoning, announced at Google I/O 2026 and expected in July.