AI Agent Astra Solves 10 Scientific Problems for Just $2,000
daniel_mac8 · x · 2026-08-02
The AI agent Astra achieved 10 scientific breakthroughs for a total cost of $2,000, successfully tackling problems that human scientists had been unable to solve.
Meanwhile, OpenAI has found major test-time compute efficiency improvements, reducing the costs of the Luna and Terra models by 80% and 20% respectively. This highlights that pushing test-time compute can significantly extend AI capabilities and cost-efficiency.
Related event: Google's Astra Solves 10 Scientific Problems for Under $2,000(4 posts)→
More from Models
- GLM 5.5 rumored to drop soon, coding ability expected to match frontier models — bindureddy · 2026-10-16
- Five rounds of adversarial prompting make Opus 5.5 and Argon converge on identical output — LanceCharger · 2026-10-16
- DIY benchmark ranks 18 models on subjective judgments, GLM-5.3 tops consensus match — arkuto · 2026-10-16
- Bindu Reddy: Google's Argon excels at complex doc writing at half the cost of Opus — bindureddy · 2026-10-16
- MAI-2 is 2x bigger than V4-Pro while V4.1F is 3x smaller: Microsoft's scaling ladder questioned — teortaxesTex · 2026-10-16
- Step 5 Preview hits DeepInfra day-zero: 600B MoE, 1M context, $1/M input tokens — StepFun_ai · 2026-10-16