FOUNDED 2024 · ATHENS · GLOBAL
About ATOM
The Global Price Benchmark for AI Inference
Independent pricing intelligence for developers, analysts, and infrastructure buyers. Transparent methodology, deterministic indexing, weekly market intelligence across the global AI inference market.
OUR MISSION
Bringing Price Discipline to Inference
Why AI pricing intelligence matters
"Markets without pricing benchmarks always overpay, always misallocate, and always consolidate around whoever sets the price first."
AI inference is the fastest-growing market in technology, and it has been operating without a price benchmark. ATOM exists to change that. We build independent, transparent, normalized benchmarks that bring price discipline to the inference economy. Built on financial index methodology comparable to S&P Dow Jones, MSCI, and Bloomberg.
OUR PLATFORM
Three Layered Capabilities, One Intelligence Stack
How ATOM is architected
Access
Three channels for delivering market intelligence to different professional contexts.
MCP
Terminal
Feed
Weekly coverage
API access
Intelligence
Analytical engine that transforms raw pricing into structured, queryable market knowledge.
Taxonomy
Classification
Normalization
Benchmark indexes
Market KPIs
Volatility composites
Vendor analysis
Model analysis
SKU analysis
Competitive rankings
Indexing
Data infrastructure that tracks the global inference market with zero variance.
Automated extraction
Change detection
Vendor monitoring
Price verification
Multi-modality coverage
Deterministic parsing
COVERAGE AT A GLANCE
The Numbers Behind ATOM
Live counts pulled from production
46,457
Datapoints
4,577
SKUs
2,962
Models
51
Vendors
35
Modalities
9
Countries
OUR INDEXES & KPIs
25 Measures Across the Platform
1 composite volatility, 15 price benchmarks, 9 market KPIs
— COMPOSITE
AIPI COMPOSITE · 1
Inference Price Volatility
The flagship cross-modality index measuring price movement across all tracked SKUs. A single number representing the global state of inference pricing, updated continuously as the underlying data refreshes.
— 15 PRICE BENCHMARKS
CHANNEL · 3
Pricing Channels
Input, Cached Input, Output. The three pricing dimensions every model exposes, indexed independently.
MODALITY · 5
Per-Format Indexes
Text, Image, Audio, Video, Embeddings. One benchmark per modality, isolating format-specific market dynamics.
TIER · 5
Market Tiers
Budget, Mid-tier, Frontier, Reasoning, Specialized. Vertical segmentation by capability and price point.
SPECIAL · 2
Cross-Cutting Benchmarks
Open-source vs proprietary pricing differentials, plus other cross-cutting indexes that span modality and channel boundaries.
— 9 MARKET KPIs
COST METRICS
Output Premium · Caching Discount · OSS Discount
Three cost-structure metrics derived from live pricing data, measuring the fundamental economics of inference workloads.
STRUCTURE METRICS
Context Cost · Model Size Spread · Reasoning Premium
Three structural metrics measuring market hierarchy: how price scales with context window, model size, and reasoning capability.
CHANNEL ECONOMICS
Platform Discount · Neocloud Discount · Caching Availability
Three channel-economics metrics measuring the arbitrage between distribution pathways and the pricing delta between providers.
FROM THE FOUNDER
"I spent 25 years helping technology companies figure out how to price their products and services. Then I realized the fastest-growing infrastructure market in the world had no pricing benchmark of its own. Markets without pricing benchmarks always overpay, always misallocate, and always consolidate around whoever sets the price first."
OUR PRINCIPLES
What Guides Every Index
Three principles, no exceptions
Objectivity
No sponsorships. No pay-to-play. Every vendor is indexed the same way, with the same rigor.
Precision
Deterministic extraction. Every SKU is normalized and verified, ensuring comparable units of measurement.
Transparency
Our methodology is public. Our data sources are documented. The index speaks for itself.
What ATOM does not do. No vendor sponsorships. No pay-to-play index inclusion. No commercial relationships influencing methodology.
WHO IT'S FOR
Built for Professionals Who Price AI
Developers, analysts, engineers
ATOM MCP
Learn more →
Developers & AI Agents
Compare GPT-4o vs Claude Opus pricing inline as you switch models. Live inference data inside Claude, Cursor, and Windsurf.
ATOM TERMINAL
Learn more →
Analysts & FinOps Teams
Track which vendor cut prices this week, see how your spend benchmarks against the market, surface arbitrage opportunities.
ATOM FEED
Learn more →
Engineering & Quant Teams
Build cost-aware routing logic with SKU-level fields. Pipe pricing data directly into your inference orchestration.
