OpenAI’s GPT-6 Astra has taken an early lead over Anthropic’s Claude Codex in internal benchmarks tailored for Indian business workflows—particularly invoice processing, GST reconciliation, and multilingual customer support. Test results circulating from enterprise pilots show Astra completing complex automations 32% faster with fewer agent handoffs, solidifying its position as the default for high-volume operational tasks.
⚡ Fast Takeaways:
- Core Update: GPT-6 Astra outperforms Claude Codex on India-specific automation benchmarks, including Tally ERP integration and Hinglish NLP tasks.
- Key Metrics / Specs: Reports indicate a 32% faster task completion rate and a 41% reduction in token spend for multi-step agentic workflows.
- Access & Availability: Rolling out to ChatGPT Team and Enterprise users now; API access expands via OpenAI’s platform to all Indian regions this week.