35B mixture-of-experts open-weights agentic model for engineering, long-horizon search and scientific research.
Updated EU status checked Last shown
Sources: creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Sage-A1.5 is a 35B Mixture-of-Experts agentic model from Alpha Innovator, building on the earlier Agents-A1. According to a YouTube creator's summary, it is based on Qwen 3.6 and tuned for scientific research. The project page covers software engineering, long-horizon search, scientific research, instruction following and tool calling.
It adds two agentic modes. Agentic Orchestration runs parallel investigations in subagents with independent contexts, which check each other's assumptions and merge their evidence. Agentic Verification is an opt-in self-review step inside the model's own chain of thought, with no external critic. The page shows demos for deep research with cited reports and for scientific tasks in neuroscience, computational chemistry, condensed matter physics, electrochemistry and microscopy image analysis.
Training has three stages: full-domain supervised fine-tuning, domain-level teacher models, and multi-teacher on-policy distillation. It uses a knowledge-action graph built from agent process traces. The weights are released openly (a creator reports about 72 GB in full and a 22 GB Q4 GGUF). It can be served with SGLang or vLLM through OpenAI-compatible endpoints, with a 262,144-token context length in the example commands. The page does not state a license or a release date, and the benchmark figures were not available in the scraped text.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Open 35B mixture-of-experts model based on Qwen 3.6, optimized for scientific research with multi-agent evidence gathering and cited reports. Beats similar-sized models on benchmarks; weights and code released (full ~72 GB, Q4 GGUF 22 GB).
Closest neighbours by meaning, not just by category.
Model
35B Mixture-of-Experts open-weights model from InternScience for long-horizon agentic search, engineering and tool use.
Model
Open-weights 35B-A3B agentic model post-trained from Qwen3.6-35B-A3B for search, coding and scientific research.
Model · Meituan
Meituan's open-weights 1.6T-parameter mixture-of-experts LLM (48B active), aimed at agentic coding and reasoning tasks.
Model
Open model family (9B dense, 35B MoE, 397B MoE) trained with a self-improvement loop for reasoning, coding and agentic tasks.
Model · Meta
Meta's open-weight 816B-parameter mixture-of-experts model (42B active, 1M context) in the Muse model family.
Model · Meta
Meta's open-weights 30B dense model for local agents, tuned for tool use and long tasks, Apache 2.0 licensed.