Astra Draws TikZ Unicorn While Scoring Near-100% on Frontier Math, ARC-AGI 3, ExploitBench
kevinweil · x · 2026-09-04
OpenAI's Sebastien Bubeck showed off Astra drawing a unicorn in TikZ, noting that a single entity can do this while also scoring essentially 100% on Frontier Math Tier 4, ARC-AGI 3, and ExploitBench — and anyone can go talk to the model directly. "Impossible to overstate the implications," he wrote.
OpenAI CPO Kevin Weil quote-shared, jokingly asking for Astra's quantitative score on the "unicorn benchmark." TikZ drawing has become a community proxy for fine-grained spatial and code reasoning ability.
Related event: Microsoft's Bubeck Shows GPT-6 Astra Drawing TikZ Unicorn(2 posts)→
More from Models
- X trends: SophontAI's Holon-S1 predicts gene activity from cancer slides; decision models rise — iScienceLuvr · 2026-10-21
- OpenAI's MultiLangBench: PHP scores highest (91.3%) at building a database engine from scratch — ricklamers · 2026-10-21
- 704-vote robotics Elo ranks Claude Opus 5.5 first at 1379; no model completes the task — qinzytech · 2026-10-21
- DeepSeek V4 called 'god-tier infra, 4th-tier ML': 7 tricks stacked, 20 patches to fix them — teortaxesTex · 2026-10-21
- GPT-6 Astra and Claude Fable 5.1 write $10,000 research papers — experts reject all — sayashk · 2026-10-21
- GitHub Copilot's HydraFusion Orchestrates Multiple Models Per Task — WirelessLife · 2026-10-21