Agnes AI
Category: Foundation Models / LLMs
Singapore-based model-first AI company building cost-efficient, SEA-focused LLMs and agentic AI infrastructure, powering the Agnes AI consumer platform. Agnes AI was founded in 2025. The company is led by Bruce Yang. Based in Singapore. Team size: 11-50. Total funding raised: $20M. Latest round: Series A. Key investors include LOOK FORWARD VCC.
- Founded
- 2025
- Headquarters
- Singapore
- Team size
- 11-50
- Total funding
- $20M
Value proposition
Delivers ~90% of top-tier AI capabilities at ~10% of the cost, targeting the "other 99.5%" of internet users underserved by premium AI platforms, with AI inclusion, parity, and neutrality as core principles.
Products and solutions
Agnes AI platform (chat, research, AI slides, AI design, group chat, filters, explore), Agnes, Echo, and Pavo applications, Agnes-1.5-Pro and Agnes-1.5-Lite 'Claw' model series, Agnes 2.5 Pro Alpha (free API on Artificial Analysis), open-source 8B-parameter SEA-focused LLM, free omni-modal AI API gateway.
Unique value
Singapore-owned AI platform independent of US and China tech blocs; positions as a potential third alternative to OpenAI and MiniMax; cost-efficient model clusters with intelligent routing delivering near-SOTA performance at a fraction of cost.
Target customer
Long-tail consumers in emerging markets (Southeast Asia, India, Middle East) plus developers/enterprises via API services; Android-first users who have never used paid AI features.
Industries served
Consumer productivity, AI infrastructure, developer API services, multimodal generation (text, image, video)
Technology advantage
RLAF (Reinforcement Learning from Agentic Feedback) framework with proprietary DSPO (Dynamic-filter Sequence-level Policy Optimization) and UV (Universal Verifier); clustered model architecture with intelligent routing and system-level orchestration; supports minority SEA languages ignored by global platforms; reasoning model using 72% fewer tokens.
How they differentiate
Model-first approach with cost efficiency (~90% capability at ~10% cost), SEA-focused language support, AI neutrality (independent of US/China blocs), and targeting long-tail consumers rather than enterprise/premium users.
Main competitors
OpenAI, MiniMax, DeepSeek, GLM (Zhipu)
Key partnerships
Google Cloud (Gemini models via Vertex AI, Cloud SQL, GKE), research collaboration with NUS and NTU co-authors.
Notable customers
Millions of users across Southeast Asia, India, Middle East, Japan, UK, and US, developer/enterprise API customers (launching Q2).
Major milestones
April 14, 2025: Agnes launched at NUS GRIP event, July 2025: official product launch on Product Hunt, 2M users in 4 months, 3M users by Dec 2025, 5M users in six months, open-sourced 8B SEA-focused LLM, Agnes 2.5 Pro Alpha debuted on Artificial Analysis with free API access, $10M Series A closed Feb 2026, approaching $20M ARR.
Growth metrics
5M+ registered users within six months of launch; ~200K-300K DAU; approaching $20M ARR; 3M users by Dec 2025.
Market positioning
Among fastest-growing AI platforms emerging from Southeast Asia; ranked in global top 10 AI labs alongside Anthropic, OpenAI, Xiaomi, and DeepSeek on benchmark; approaching $20M ARR; working toward potential SGX listing by end of 2026.
Geographic focus
Southeast Asia (primary), India, Middle East, Japan, UK, US
Patents and IP
Proprietary research systems: RLAF framework, DSPO (Dynamic-filter Sequence-level Policy Optimization), UV (Universal Verifier); open-sourced 8B SEA-focused LLM on GitHub (AgnesAI-Labs).
About Bruce Yang
UC Berkeley dual degrees in Mathematics and Computer Science; worked at Microsoft and LinkedIn; co-founded a startup that achieved millions of downloads; PhD candidate in AI at National University of Singapore (NUS); Raffles Institution alumnus
Latest news about Agnes AI
More Foundation Models / LLMs companies
Official website: https://www.agnes-ai.com/