AXIOM
Get API Access
Frontier AI Research
Vol. I, Issue 1
Est. 2024

01αβΓΔπΩ∂∇∑∫

Frontier research. Radical safety. Constitutional foundations.

Mission

Intelligence aligned with humanity’s future.

We believe that the development of advanced AI is the most consequential technological transition in human history — and that getting it right requires scientific rigor, philosophical depth, and an unflinching commitment to safety.

Learn our approach
Models

Frontier capabilities,
safety-first foundations.

View all models
LatestAPI Available

AXIOM 2

Our most capable and safest model

State-of-the-art performance on scientific reasoning, mathematics, and code — built with safety-first training from the ground up. Every capability advance is matched by rigorous safety evaluations.

200Ktoken context
  • Advanced scientific reasoning
  • Complex mathematical proofs
  • Production-grade code generation
  • Nuanced long-form analysis
Research PreviewAPI Available

AXIOM Research

Frontier capabilities for research workloads

1.0Mtoken context
  • 1M token context window
  • Extended reasoning mode
  • Research API access
FastAPI Available

AXIOM Instant

Production speed for everyday tasks

32Ktoken context
  • Sub-100ms response times
  • Competitive reasoning ability
  • High throughput API
ResearchAll papers

Advancing the science of safe AI.

December 2025

Constitutional AI: Harmlessness from AI Feedback

We introduce a method for training a harmless AI assistant through self-improvement, without any human labels identifying harmful outputs. The core idea is to use a list of rules about harmlessness—a constitution—to automatically correct outputs. We train both a supervised policy and a reinforcement learned policy, and find that the RL policy achieves better results than RLHF while generating fewer harmful outputs.

Dr. Sarah Chen · Marcus Webb · Dr. Aisha Patel · James Liu · Dr. Elena Vasquez

SafetyRLHFConstitutional AIAlignment
Read paper

"We introduce a method for training a harmless AI assistant through self-improvement, without any human labels identifying harmful outputs. The core idea is to use a list of rules about harmlessness—a ..."

September 2025

Toy Models of Superposition

Neural networks often represent more features than they have dimensions. We call this phenomenon superposition and explore it in toy models, showing how features can be packed into fewer dimensions through interference patterns. Understanding superposition is crucial for mechanistic interpretability.

Dr. Aisha Patel · Dr. James Liu · Dr. Elena Vasquez

InterpretabilityMechanistic AnalysisNeural NetworksFeatures
Read paper

"Neural networks often represent more features than they have dimensions. We call this phenomenon superposition and explore it in toy models, showing how features can be packed into fewer dimensions th..."

2.4T
Parameters
Across all model variants
183+
Research Papers
Published since 2024
52
Languages
Supported natively
99.97%
API Uptime
30-day rolling average
Testimonials

From researchers and engineers at leading institutions

Trusted by the teams\npushing the frontier.

AXIOM 2 is the first model I trust with sensitive scientific analysis. Their safety commitments aren't marketing — they're reflected in every interaction.

Prof. David Kim
Director of AI Research · MIT CSAIL

Constitutional AI is a genuine breakthrough. The model refuses harmful requests without becoming uselessly cautious. That balance is extremely hard to achieve.

Dr. Priya Mehta
Senior Researcher · Oxford Future of Humanity Institute

We evaluated six frontier models. AXIOM 2 was the only one that consistently declined to assist with CBRN tasks while maintaining full capability on legitimate research.

Thomas Ellison
AI Safety Lead · UK AI Safety Institute

The 200K context window isn't just a number — it actually works. We process entire genomics papers in a single pass. The comprehension quality is remarkable.

Dr. Sofia Andersen
Computational Biology Lead · Stanford Bio-AI Lab

What impresses me most is the calibration. When AXIOM doesn't know something, it says so. That epistemic honesty is rarer than it should be in this industry.

James Nakamura
CTO · Synthesis Therapeutics

Their RSP and Constitutional AI approach gives us confidence to build on top of AXIOM. We know the foundation is safety-first, not safety-as-afterthought.

Dr. Amara Osei
Chief Science Officer · DeepCure Technologies

Trusted by the world's most ambitious research teams

Get Started

Ready to build with frontier AI?

Access AXIOM 2 through our API. Production-grade safety, world-class capabilities, transparent pricing.