Humans&
Category: Foundation Models / LLMs
A frontier AI research lab developing human-centric generative models designed to augment human intelligence and foster collaborative reasoning through high emotional intelligence (EQ). Humans& was founded in 2025. The company is led by Eric Zelikman. Based in San Francisco, USA. Team size: 10-50. Total funding raised: $480.0M. Latest round: Seed ($480.0M, Jan 2026). Key investors include SV Angel (lead), Georges Harik (lead), Nvidia, Jeff Bezos, GV (Google Ventures), Emerson Collective, Forerunner, S32, DCVC, Human Capital, Felicis, CRV.
- Founded
- 2025
- Headquarters
- San Francisco, USA
- Team size
- 10-50
- Total funding
- $480.0M
Value proposition
Moves beyond 'chatbot' interactions to 'collaborative reasoning,' providing AI that understands human intent, processes internal thought (Quiet-STAR), and prioritizes emotional intelligence to improve decision-making.
Products and solutions
Frontier Human-Centric LLMs, Collaborative Reasoning Interface, EQ-Integrated Reasoning Engine, Self-Taught Reasoning (STAR) Frameworks
Unique value
Unlike traditional LLMs that focus on next-token prediction, Humans& focuses on 'internal monologue' and reasoning-first architectures that mimic human cognitive flexibility and empathy.
Target customer
Knowledge workers, enterprise research teams, creative professionals, and organizations requiring AI that integrates into human workflows rather than replacing them.
Industries served
Artificial Intelligence, Enterprise Software, Research & Development, Professional Services, Education & Creative Arts
Technology advantage
Leverages the 'Self-Taught Reasoner' (STAR) and 'Quiet-STAR' algorithms, allowing models to learn how to reason independently and process thoughts before generating output, significantly reducing hallucinations and improving complex problem-solving.
How they differentiate
Focuses on 'Collaborative Reasoning' and Emotional Intelligence (EQ) rather than just task automation. Utilizes proprietary 'Quiet-STAR' algorithms to give AI an internal monologue, allowing it to reason before speaking, which contrasts with the standard next-token prediction models of competitors.
Main competitors
OpenAI, Anthropic, xAI
Key partnerships
NVIDIA (Strategic investor; technical collaboration on NVFP4 RL training stack with open-source contributions), RadixArk (RL systems collaboration), SV Angel (seed lead), Stanford University (academic ties via co-founder Noah Goodman)
Notable customers
Enterprise Research Teams, Frontier AI Labs, Knowledge Work Professionals
Major milestones
Company founded September 2025 by Eric Zelikman, Georges Harik, Andi Peng, Noah Goodman, Yuchen He (and founding team incl. Ray Ramadorai), Emerged from stealth Jan 20 2026 with $480M seed at $4.48B valuation led by SV Angel & Georges Harik, Published open-source NVFP4 RL training recipe (July 2026) with NVIDIA/RadixArk collaboration
Growth metrics
Achieved a $4.48B valuation at the Seed stage; team composed of elite researchers from xAI, Google, and Anthropic.
Market positioning
Frontier AI research lab positioned as a human-centric alternative to 'replacement-style' AI, targeting high-stakes enterprise reasoning and creative collaboration.
Geographic focus
Global (Headquartered in San Francisco, USA)
Patents and IP
Proprietary implementations of STAR and Quiet-STAR algorithms; specific patent filings are not yet public but are based on CEO Eric Zelikman’s pioneering research at Stanford.
About Eric Zelikman
Eric Zelikman is co-founder and CEO of humans&. Early xAI employee: key contributor to Grok 2 pretraining data; kicked off and scaled RL for reasoning on Grok 3 Thinking; built agent RL infrastructure/recipe for Grok 4. Stanford PhD candidate (on leave) advised by Nick Haber and Noah Goodman; inventor of STaR and Quiet-STaR. Focuses on human-centric AI with EQ and collaborative reasoning.
Latest news about Humans&
More Foundation Models / LLMs companies
Official website: https://humansand.ai/
