India unveils Eka-1, Dhir 2B, and Dhir 5B: three indigenous foundation AI models built for offline edge inference, vernacular token efficiency, and sovereign national security.
OpenAI has released AI-generated research results across three fields, while inviting outside experts to scrutinize and explain the findings.
Cross-reference verified evaluation benchmarks (AIME, Indic MMLU, SWE-bench) and estimate production token costs.
Live evaluation matrix tracking reasoning, Indic language fidelity, API economics, and autonomous coding.
Calculate your startup or enterprise monthly inference bill across frontier models.
AI Latest News is India's premier independent digital publication dedicated exclusively to artificial intelligence, foundation model engineering, and sovereign computing infrastructure. As global frontier labs compete on reasoning and multimodal architectures, our mission is to deliver fast, rigorous, and verified reporting tailored specifically for Indian software developers, startup founders, policy makers, and researchers.
Our reporting is organized into six specialized desks: breaking global developments in AI News; architecture and parameter teardowns in Model Launches; vetted developer workflows and productivity software in AI Tools; in-depth coverage of the ₹10,372 crore IndiaAI Mission and domestic models like Sarvam and Krutrim in India AI; rigorous head-to-head benchmarks in Guides and Comparisons; and production deployment architectures in Enterprise AI. All claims and benchmarks are backed by primary source documentation and reproducible tests.
Common questions about AI Latest News, Indic models, and sovereign compute in India