Murf AI Introduces Falcon 2 Voice Model

Murf AI announced Falcon 2, a text-to-speech model scheduled for public availability on August 20 at $0.01 per generated minute. Bloomberg reports that the Bangalore-based startup's model ranked above some OpenAI systems, including the Realtime API, on Artificial Analysis. Murf's pricing, latency, and enterprise deployment claims place the release in an increasingly competitive synthetic-speech market.
Murf AI has announced Falcon 2, a text-to-speech model scheduled to become publicly available on August 20 at $0.01 per generated minute. Bloomberg reports that the Bangalore-based startup's model ranked above some systems from larger providers, including OpenAI's Realtime API, on the Artificial Analysis performance-tracking platform.
The release puts Murf in direct competition with a crowded set of voice-model providers, including OpenAI, ElevenLabs, and other API vendors. The Next Web reports that Murf is pricing Falcon 2 at one-fifth of the cited cost of ElevenLabs Turbo and Flash, equating to $10,000 for one million generated minutes under the stated rate.
Benchmark and performance claims
Murf has cited the independent Artificial Analysis leaderboard in its performance comparisons. Bloomberg reported that Falcon 2 ranked higher than some OpenAI platforms, while The Next Web reported that Murf claimed a score above some OpenAI entries, including the Realtime API, and response times below 100 milliseconds.
Artificial Analysis lists Falcon 2 with a Quality Elo of 1154.18 and places it 29th among 98 models in its Voice Arena quality ranking at the time of retrieval. The platform describes its Elo metric as an average rating based on user responses, with higher scores indicating stronger perceived quality. That metric is useful comparative evidence, but it is not equivalent to an independently audited production evaluation across every language, accent, or deployment setting.
Analytics India Magazine reports that Falcon 2 supports more than 150 voices across 35 languages and is designed for multilingual conversations and varied accents in the Indian market. The publication also reports that Murf intends to price Falcon 1 and Falcon 2 similarly and teased a possible Falcon 3 release in the coming months.
Cost, latency, and deployment
The Next Web reports that Falcon 2 includes data residency options in 11 countries, customer-cloud deployment, and support for up to 10,000 concurrent calls. It also reports that Murf works with voice actors under consent-based contracts that include ongoing royalties and data protections, citing chief operating officer Sneha Roy.
For teams deploying text-to-speech systems, price per generated minute is only one part of the operating-cost calculation. Comparable deployments typically need to evaluate time-to-first-audio, sustained throughput, multilingual naturalness, interruption handling for real-time agents, voice consistency, and controls around speaker likeness. A leaderboard score can help narrow a vendor shortlist, but application-specific listening tests and load tests remain necessary for production selection.
Murf launched the first Falcon model in November 2025, according to The Next Web. The publication also reports that Murf, founded in 2020 by three IIT Kharagpur engineers, claims 10 million users and 300 corporate customers, including Cisco, Pfizer, Nestle, and Air France-KLM. Those customer and user figures are company claims reported by The Next Web.
Falcon 2's release reflects a broader pattern in speech AI: providers increasingly compete on a combination of subjective voice quality, low-latency streaming, governance controls, and unit economics. For ML and product teams, the relevant comparison is therefore likely to extend beyond a single naturalness ranking to the full inference and integration profile of each voice API.
Key Points
- 1Murf AI's Falcon 2 enters public availability at $0.01 per generated minute, making voice inference pricing a central competitive variable.
- 2Artificial Analysis lists Falcon 2 at 1154.18 Quality Elo and 29th of 98 models, offering independent but limited comparative evidence.
- 3Comparable production voice deployments typically require application-specific tests for latency, multilingual quality, throughput, and speaker-governance controls.
Scoring Rationale
Falcon 2 is a notable text-to-speech release with aggressive stated pricing and independently listed leaderboard data. It is relevant to teams selecting voice APIs, especially for real-time and multilingual deployments, but its broader production performance remains to be established.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems

