Agents & Benchmarks
Sentient has released multiple open-source agents that have shown strong benchmark performance:
ROMA
• Open-source meta-agent orchestration framework
• Reported to outperform systems from Perplexity, OpenAI, and Gemini on research benchmarks
• GitHub traction: ~2.7k stars, ~265 forks
• ROMA V2 adds auto-review, reasoning upgrades, and enterprise-ready deployment
SERA
• Open-source AI agent designed for crypto and DeFi research
• Ranked #1 on Web3 & DeFi agent benchmarks
• Benchmarked ahead of GPT-5 Medium Reasoning, Grok 4, and Perplexity Finance
These results suggest open-