Prediction may be the hardest benchmark in AI. No memorization. No retrieval. No leakage. No spoilers. The answer literally doesn’t exist yet. Excited to see Apodex-1.0-mini (#1 this week) performing well in that environment.

We submitted four experimental prediction harnesses built on Apodex-1.0-mini to FutureX. They ranked #1, #2, #3, and #4🚀 One 35B model. Future prediction requires discovering, verifying, and synthesizing evidence under uncertainty.

Today, we’re excited to introduce Apodex 1.0. 🚀 We believe the future of AI research isn’t about generating more answers—it’s about generating better understanding. Apodex is built for long-horizon research: planning investigations, gathering evidence, using tools, testing hypotheses, and synthesizing grounded conclusions. The name Apodex comes from the Greek word apodeixis (ἀπόδειξις), meaning demonstration or proof. We chose it because knowledge becomes valuable only when it can be shown, supported, and understood—not merely asserted. This release includes: • Apodex-1.0-H, our flagship deep research model • AgentOS, a runtime for building and evaluating agent workflows • Apodex-1.0 and Apodex-mini • Open-weight Smol models at 0.8B, 2B, and 4B Apodex-1.0-H achieves state-of-the-art results across multiple research benchmarks: • 90.3 on BrowseComp • 60.8 on HLE-Text • 94.4 on DeepSearchQA • 46.7 on FrontierScience Research • 87.4 on FrontierScience-Olympiad • 74.2 on SuperChem We’re excited to share Apodex with researchers, scientists, developers, and builders worldwide. 🔗 Learn more: Blog: lnkd.in/gDnaj2cn Tech Report: lnkd.in/gSXEuUfi GitHub: lnkd.in/g5uPhJyM Hugging Face: lnkd.in/ggkhspcV Discord: lnkd.in/ghPED8v9 API Platform: lnkd.in/grwQB-te