Apex-35B-V1.0: a 35B general-purpose model built with an automated AI research loop
Beijing startup Apex Intelligence released its first model, a 35B model built on Qwen3.6-35B-A3B via mid- and post-training, with AI doing data construction, experiment design and training optimisation in a recursive self-improvement loop.
Why it matters
Claims to beat its Qwen3.6 base on 20 of 21 benchmarks, with a 32.1-point CFD-Bench gain attributed to the automated loop — a small lab claiming AI-built-model gains, not just a fine-tune.
What you could build with it
A computational science lab could fine-tune the model on its own simulation workflows to draft experiment plans and analysis code, cutting the time researchers spend setting up fluid-dynamics runs.
Does it hold up?
Downloadable today with a deployment guide, but all benchmark claims are company-reported and independent evaluations have not yet appeared.
Built with Apex-35B-V1.0
- Apex Intelligence releases Apex-35B-V1.0, cites 32.1-point CFD-Bench gainblog · Independent writeup of the automated research system, the throughput claims, and the company-reported benchmark table.
- Apex Intelligence releases Apex-35B-V1.0, a general-purpose model built with automated AI researcharticle · Official release detailing the recursive self-improvement loop and availability on Hugging Face and GitHub.
- apexin-ai/Apex-35B-V1.0github · Official repository with weights, deployment guide and usage examples.
Learn more
First spotted on hf: source.
More AI models
Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73NVIDIA Isaac Lab 3.0 GA + Isaac OS open runtime + Agile ONE SN humanoidMajor Isaac platform expansion at GTC Berlin Oct 21: Isaac Lab 3.0 is generally available as the open-source robot…models · JEV 0.73Kimi K3.1 Pro: Moonshot's 2.8T open-weight flagship officially launchesMoonshot AI officially launched Kimi K3.1 Pro on Oct 23, a mixture-of-experts model with 2.8 trillion total / 104…models · JEV 0.72NVIDIA Nemotron 3.5 Retriever collection: Embed 8B/1B, Rerank 8B, Parse 631MNVIDIA's Oct 21 collection drop for agentic retrieval: Nemotron-3.5-Embed in 8B BF16, 1B BF16 and 1B NVFP4…models · JEV 0.72NVIDIA Cosmos 3.5 world models: now openly availableNVIDIA's Cosmos 3.5 family of open world foundation models for robots, autonomous vehicles and vision agents is now…models · JEV 0.71NVIDIA RFD4-Proteina: open protein-design BioNeMo modelOpen protein-design model from NVIDIA and the Baker Lab at UW Medicine's Institute for Protein Design, announced at…models · JEV 0.7
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.