AI

Meta sets Indian subscription rates for Muse AI agent ahead of launch

Power tier priced at ₹1,799 per month as tech giant rolls out lightweight Muse Ember model to slash enterprise agent execution costs.

By KaroStartup News Engine

Meta has set Indian subscription prices for its Muse AI agent ahead of its official launch in the country, introducing a Power tier at ₹1,799 per month and a Maximum tier at ₹8,999 per month. Subscriptions bought through Google Play and Apple billing systems will incur slightly higher fees. The pricing structure, which mirrors US rates, comes as Meta Superintelligence Lab also released Muse Ember, a lightweight model priced at $0.10 per million input tokens designed to run repetitive agent tasks at scale.

Muse pricing and distribution in India

The subscription rates put Meta's flagship agentic AI directly into the Indian market, where Muse has already gained traction despite operating under a waitlist. Users in India have accessed the platform by registering through virtual private networks connected to supported markets including the US, Canada, Japan, and Singapore. Once registered, accounts function over Indian internet connections, with web and WhatsApp interfaces accessible locally. Indian-issued credit and debit cards can process subscription payments provided the billing address is set to a supported market.

Full local functionality remains blocked by technical and operational hurdles. Muse currently cannot accept user authorisation to perform autonomous card payments on their behalf in India, as the service lacks integrations with local payment gateways. An executive at a major Indian payment provider confirmed last month that the firm is not collaborating with Meta's developers to build agentic payment capabilities in the region.

Muse reached over 10 million downloads globally within two months of its September release. Its primary distribution advantage in India rests on its WhatsApp integration, allowing users to initiate tasks inside the messaging application. Main competitor Instinct, a San Francisco-based agentic platform, also operates through WhatsApp but remains available globally without a paid subscription tier.

Muse Ember model and benchmark performance

Alongside the India pricing reveal, Meta Superintelligence Lab introduced Muse Ember, its smallest and most cost-effective execution model to date. Built specifically for token-heavy, repetitive agentic workloads such as tool calls, document parsing, routing, and coding subtasks, Ember is priced at $0.10 per million input tokens, $0.004 per million cached input tokens, and $0.50 per million output tokens on the Meta Model API.

Meta CEO Mark Zuckerberg described Ember as "our smallest model that delivers higher intelligence per dollar than others in its class," adding that it was built for tasks agents repeat continuously. Meta AI chief Alexandr Wang noted that the model provides "more intelligence per dollar than Haiku, Luna, and all other flash models at any reasoning setting and does it at a third of the cost or less."

Ember is designed to work in tandem with Muse Wonder, the larger frontier model Meta released in October. Under this two-tier architecture, Muse Wonder establishes high-level instructions and planning, while Muse Ember executes high-volume operational tasks across text, image, audio, and video inputs.

On the Artificial Analysis Intelligence Index, Muse Ember running at maximum reasoning effort achieved a composite score of 35.8 at an estimated cost of $0.136 per task. This puts it nearly level with Anthropic's Claude Haiku 5.5 at max reasoning, which scored 36.2 at $0.411 per task, while outperforming OpenAI's GPT-6 Luna max score of 27.8 at $0.152 per task.

On workflow automation tests, Ember scored 37.8 on AutomationBench-AA, compared to 13.9 for Haiku 5.5 and 20.2 for Luna. Haiku 5.5 retained a lead in engineering code merges, scoring 34.9 on FrontierCode 2.0-AA against Ember's 27.2.

Why it matters for Indian developers and enterprises

For Indian founders, software engineers, and enterprise technology operators, Meta's twin announcements mark a sharp reduction in the unit economics of deploying autonomous agents. Running production-level AI agents previously required choosing between expensive frontier models or lower-grade execution tools. At $0.10 per million input tokens, Ember presents an alternative to open-source flash models like DeepSeek V4.1 Flash and GLM 5.3 Flash, enabling Indian engineering teams to run high-frequency tasks such as invoice parsing, code checking, and customer interaction routing at lower overheads.

However, enterprise adoption in India faces regulatory and technical constraints. The inability of Muse to execute card payments locally restricts fully automated transaction workflows. Furthermore, autonomous agents capable of bypassing standard CAPTCHA and legacy bot-detection mechanisms present security and compliance questions for Indian digital infrastructure and corporate networks.

The Muse family expansion

Muse Ember represents the fifth major release under Meta's Muse product brand within seven months. The family expanded rapidly through 2026:

What happens next

Meta has not specified an official date for launching Muse directly on Indian app stores or finalizing rupee billing structures. As competition in autonomous AI execution sharpens between Meta, Anthropic, OpenAI, and standalone platforms like Instinct, Indian payment gateways and digital service providers are expected to face growing pressure to support agentic integration protocols.