Murf AI, a Bengaluru-based startup, has launched Falcon 2, a text-to-speech model that outperforms ElevenLabs and OpenAI on independent quality benchmarks while charging roughly $0.01 per minute of generated speech, compared to ElevenLabs' $0.05 per minute. The model generates audio in under 100 milliseconds and supports over 150 voices across 35 languages, including 12 Indian languages.

How Does Falcon 2 Compare to Competitors on Quality and Speed?

Falcon 2 ranks above ElevenLabs' Flash v2.5 and Turbo v2.5 voices, OpenAI's Realtime API, and xAI's text-to-speech offering on Artificial Analysis's naturalness scoring, an independent, listener-voted benchmark that the voice AI industry treats as its standard leaderboard. The model achieves this while maintaining latency under 100 milliseconds, meaning it can start generating audio fast enough that a phone conversation doesn't feel like talking to a machine on a satellite delay.

It's important to note that leaderboard rankings on Artificial Analysis shift frequently as labs release updates, so "ranks above X and Y today" represents a snapshot rather than a permanent title. Additionally, naturalness is just one performance axis; how the model performs on specific accents, scripts full of policy numbers, or noisy call-center environments requires real-world testing.

What Makes the Pricing Gap Significant for Enterprises?

The cost difference becomes substantial at scale. At Murf's rate of $0.01 per minute, generating one million minutes of speech costs $10,000. The same volume costs $50,000 at ElevenLabs' $0.05-per-minute rate. For a mid-sized outbound-calling or interactive voice response (IVR) operation, this represents a savings of roughly 38 lakh rupees (approximately $40,000 USD) per million minutes generated, a volume that mid-sized operations could plausibly hit within a few months.

Murf's engineering approach prioritizes efficiency without relying heavily on high-end hardware. The company uses smaller, commodity graphics processing units (GPUs) rather than expensive specialized chips, allowing it to manage infrastructure costs as demand fluctuates. Murf also offers on-premise deployment, allowing customers in regulated industries like banking and financial services to run a synthesized version of the model on their own servers.

Who Is Murf AI and What's Their Track Record?

Murf AI was founded in October 2020 by three IIT Kharagpur classmates: Sneha Roy, Ankur Edkie, and Divyanshu Pandey. The company started as a voiceover tool for presentations and video before pivoting toward the real-time, developer-facing API business that Falcon and Falcon 2 represent. Murf has raised $11.5 million total: a $1.5 million seed round led by Elevation Capital in 2021, followed by a $10 million Series A led by Matrix Partners India in 2022. This is a modest war chest compared to ElevenLabs, which has raised over $100 million.

"For years, the most important AI models have largely emerged from the US," said Sneha Roy, COO at Murf AI.

Sneha Roy, COO at Murf AI

Falcon 2 is the second generation of Murf's Falcon line, which first launched in November 2025 with claimed 55-millisecond model latency. Falcon 1 was in limited release to select enterprise clients before Falcon 2's broader public availability, which is standard practice for a model this latency-sensitive. According to Bloomberg's reporting, Falcon is already being used in production by Honeywell, Cisco, Pfizer, VMware, Nestlé, and Air France-KLM.

Why Does This Matter for India's Business Process Outsourcing and Financial Services?

India's call-center and business process outsourcing (BPO) industry represents the exact buyer Falcon 2 is engineered for: high call volume, thin profit margins, and genuine appetite to automate reminder calls, collections, know-your-customer (KYC) verification prompts, and basic customer queries without sounding like an outdated IVR system. At $10,000 per million minutes versus $50,000, the price gap alone could determine whether a voice AI pilot gets funded or shelved after the first cost review.

Murf ran a public "10 Days of Voice Agents" builder challenge between August 6 and 15, tasking developers with building a voice agent capable of handling a real phone call for someone in India by Independence Day, using Falcon 2 as the underlying engine. This signals Murf's positioning of Falcon 2 as infrastructure for the Indian voice-agent developer ecosystem, not just an enterprise sales pitch aimed at Fortune 500 logos.

Steps to Evaluate Falcon 2 for Your Organization

• Test on Your Own Data: Benchmark naturalness scores are measured on standardized prompts, not on your specific accent mix, compliance-heavy scripts, or background-noise conditions. Run a pilot before committing to a full deployment.
• Calculate Your Volume Economics: Determine how many minutes of generated speech your organization will need monthly or annually. Use the $0.01-per-minute rate to project costs and compare against your current solution or competitors' pricing.
• Assess Language and Voice Requirements: Falcon 2 supports over 150 voices across 35 languages, including 12 Indian languages. Verify that the specific language variants, accents, and voice characteristics you need are available.
• Evaluate Deployment Options: Decide whether cloud-based API access meets your needs or whether on-premise deployment is required for regulatory or security reasons.

Falcon 2 isn't launching into empty space. Sarvam AI's Bulbul model and the broader IndiaAI Mission-backed sovereign AI push are also targeting Indian-language voice interfaces, and global players like Cartesia have been adding Indian-language support to their own leaderboard-topping models. Falcon 2's pitch is less about being India-specific in language coverage and more about being India-priced at global quality, a different and arguably more immediately commercial bet.

"Improving voice quality was the main focus of Falcon 2, while the company retained the speed and cost profile of its earlier model," said Ankur Edkie, co-founder and CEO at Murf AI.

Ankur Edkie, Co-founder and CEO at Murf AI

Murf has completed a Series A and is considering a Series B, with discussions likely to begin toward the end of the year. An initial public offering is not an immediate priority as Murf continues to expand its product offering and build out the broader voice-agent stack.