About
FriendliAI is a generative AI infrastructure company offering the fastest inference platform for deploying and scaling AI models. Their purpose-built stack delivers 2x+ faster inference with ultra-low latency and cost efficiency, serving businesses that need to turn latency into a competitive advantage. They provide guaranteed reliability with 99.99% uptime SLAs and seamless deployment of over 510,000 Hugging Face models or custom proprietary models.
Key facts
At-a-glance profile data
- Website
- friendli.ai
- Headquarters
- Redwood City, United States
- Founded
- 2021
- Employees
- 53
- Total funding
- $66M (Series A)
- Business model
- B2B
FriendliAI pricing
| Product | Plan | Price | Includes |
|---|---|---|---|
| Model APIs | GLM-5.3 | $1.26 per 1M input tokens, $0.234 per 1M / 1M tokens | 10% OFF listed price, regular $1.4 input, $0.26 cached input, $4.4 output |
| Model APIs | GLM-5.3-Flash | $0.15 per 1M input tokens, $0.03 per 1M / 1M tokens | - |
| Model APIs | DeepSeek-V3.2 | $0.5 per 1M input tokens, $0.25 per 1M c / 1M tokens | - |
| Model APIs | MiniMax-M2.5 | $0.3 per 1M input tokens, $0.06 per 1M c / 1M tokens | - |
| Model APIs | gemma-4-31B-it | $0.14 per 1M input tokens, $0.4 per 1M o / 1M tokens | - |
| Model APIs | whisper-large-v3 | $0.0015 per audio minute / audio minute | Speech to text pay per second of audio input |
| Dedicated Endpoints | A100 80GB | $2.9 per hour / hour | Billed per second, $4.0 per hour effective Oct 1 |
| Dedicated Endpoints | H100 80GB | $3.9 per hour / hour | Billed per second, $5.0 per hour effective Oct 1 |
| Dedicated Endpoints | H200 141GB | $4.5 per hour / hour | Billed per second, $7.0 per hour effective Oct 1 |
| Dedicated Endpoints | B200 180GB | $8.9 per hour / hour | Billed per second, $9.0 per hour effective Oct 1 |
| Dedicated Endpoints | B300 288GB | $12.0 per hour / hour | Billed per second, $12.0 per hour effective Oct 1 |
Text and vision models pay per token. For models where cached input pricing is not listed, prompt caching discounts may be available for enterprise deployments. Contact sales for details.
On-demand deployment only pay for compute you use, down to the second, with no extra charges for start-up times. Pricing through Sep 30 and effective Oct 1 as listed.
Plans as published by FriendliAI, checked September 2026. Current prices on friendli.ai
Funding rounds
| Date | Round | Amount | Lead investors |
|---|---|---|---|
| 29 Aug 2025 | Series A | $20M | Lightspeed Venture Partners, Matt Bornstein, Capstone Partners |
| 4 Jun 2024 | Undisclosed round | $20M | - |
| 23 May 2023 | Undisclosed round | $20M | - |
| 1 Jan 2021 | Seed | $6M | Capstone Partners |
Tech stack
Detected on friendli.ai, last scanned 21 Sept 2026. Run your own scan
- Frontend
- Next.js, Tailwind CSS
- Analytics
- Google Tag Manager
- CDN
- Vercel
- Email provider
- Google Workspace
- Email marketing
- Customer.io
- Status page
- Instatus
Frequently asked
How much funding has FriendliAI raised?
FriendliAI has raised a total of $66M in funding. The most recent round on record is Series A.
How many employees does FriendliAI have?
FriendliAI has approximately 53 people on record.