Bengaluru-based AI startup Sarvam AI has unveiled a redesigned brand identity under the tagline “The all-new Sarvam — Designed for all of India,” bringing together its recent artificial intelligence launches under a unified platform narrative.
The brand refresh follows a week of heightened attention for the company after it claimed that its AI tools — Sarvam Vision and Bulbul V3 — outperform global competitors such as Google Gemini and OpenAI’s ChatGPT on select benchmarks.
Sarvam AI said the refreshed identity reflects its long-term vision of “AI for all, from India.” The name Sarvam, derived from the Sanskrit word for “all,” sits at the heart of the company’s ambition to build AI systems that reflect “how this country thinks, speaks, reasons, and solves problems,” while still meeting global standards of technical excellence.
According to the company, the rebrand signals a shift from individual product launches to a cohesive platform identity focused on India-first artificial intelligence.
The new design centres on the concept of a “gateway,” described as a threshold where “human and machine” and “culture and computation” meet.
Inspired by mandala construction, the Sarvam monogram uses geometric repetition to form a layered, lotus-like structure. A blue-to-orange gradient represents motion and progression, while the overall visual system is intended to feel “grounded, familiar, and modern at the same time.”
The redesigned identity sparked strong reactions on X, with several users praising the branding as “clean” and distinctly Indian. Commenters noted similarities to mandalas and the arched windows of Rajasthan’s forts, calling it “Indic design done right.”
At the same time, users questioned when Sarvam would introduce a consumer-facing app. Currently, the company’s products are positioned as APIs and platforms for developers, businesses, and institutions, which explains the absence of a public app for now.
The branding refresh brings together Sarvam’s recent product announcements into a single, sharper narrative focused on India-centric AI development.
The company had already drawn attention earlier this week with the launch of Sarvam Vision and Bulbul V3, both of which it claims outperform international AI models on specific benchmarks.
Sarvam Vision is the company’s document-reading and OCR model, designed to extract text from images, scanned documents, and handwritten notes.
Beyond basic OCR, the model can:
Extract data from graphs
Interpret charts and trends
Preserve complex table structures, including nested or visually complicated tables
The model supports 22 languages, including major Indian languages such as Hindi, Bengali, Tamil, Telugu, and Marathi.
According to co-founder Pratyush Kumar, Sarvam Vision achieved:
84.3% accuracy on the olmOCR-Bench, compared to
Gemini 3 Pro: 80.20%
DeepSeek OCR v2: 78.80%
ChatGPT: 69.80%
93.28% on OmniDocBench v1.5, versus
Gemini 3 Pro: 91.6%
ChatGPT 5.2: 86.56%
87.36% word accuracy, higher than
Gemini 3 Pro: 82.51%
ChatGPT 5.2: 38.60%
Sarvam AI also launched Bulbul V3, a next-generation text-to-speech model designed specifically for Indian languages.
Key features include:
35+ high-quality voices
Support for 11 Indian languages
Handling of code-mixing, regional accents, names, abbreviations, and emotional tone
As the company has noted earlier, Indian speakers often switch languages mid-sentence, making adaptability critical for voice systems.
Bulbul V3 has been tested in independent listening studies, where it scored strongly on naturalness, robustness, and stability.
The model focuses on pacing, emphasis, and emotional tone, ensuring that speech sounds natural from the first few seconds instead of mechanically generated.
The redesigned Sarvam brand ties Vision and Bulbul V3 together under a unified message: building AI systems that work reliably across India’s linguistic diversity and real-world usage patterns.
This evolution reflects the broader vision of co-founders Pratyush Kumar and Vivek Raghavan, who have consistently positioned Sarvam AI around creating scalable, India-native AI systems that meet global performance benchmarks.