Holo4: H Company's open-weight models for computer-use agents
H Company released Holo4 on Sept 28, 2026 - two open-weight vision-language models (27B dense, 35B-A3B MoE) purpose-built for computer-use agents that click, type, write code, and call MCP/API tools across web, desktop, and mobile.
Why it matters
First open-weight computer-use model family designed to act through any interface (GUI, code, MCP, APIs); the 27B model reportedly scores 85.2% on OSWorld at $0.08/task, with weights on Hugging Face in BF16, FP8, NVFP4, and GGUF.
What you could build with it
A product team building RPA or desktop-automation tooling can swap proprietary agent models for self-hosted Holo4 and own the full execution loop; an indie hacker can ship a one-click AI intern that operates a browser plus a desktop VM to handle back-office workflows; a QA team can run Holo4 against staging environments for end-to-end UI testing without paying per-task API fees.
Does it hold up?
Too early to judge - launched today; only vendor-reported OSWorld numbers exist so far, with no independent community evaluations yet.
Learn more
First spotted on article: source.
More AI models
ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69kev: open, trainable Jev-like family of small decision models on Qwen3.5/3.8A family of small decision models (0.8B to 27B) built on Qwen3.5/3.8 that reproduces Jev's prefill-only typed-decision…models · JEV 0.68GPT-Synopsys: OpenAI and Synopsys multi-year deal to build an AI model that operates EDA toolsOpenAI and Synopsys signed a multi-year agreement on Sep 30 to jointly develop GPT-Synopsys, a specialized model that…models · JEV 0.66
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.