| FEED

Inference Pricing Data, Direct to Your Systems

Name: ATOM Direct Feed
Brand: ATOM

Structured inference pricing data, direct to your systems. 1,600+ SKUs across 40 vendors. Query by 32 parameters.

GET api.a7om.com/v1/prices?model=llama-4-maverick&vendor=together-ai

{
  // ── Identity ────────────────────────────────────
  "sku_id": "together-llama-4-maverick-text-input",
  "vendor_name": "Together AI",
  "vendor_type": "Inference Platform",
  "model_name": "Llama 4 Maverick",
  "sku_plan_name": "Llama 4 Maverick — Text Input",
  // ── Model Intelligence ───────────────────────────
  "modality": "Text",
  "model_family": "Llama 4",
  "context_window": 1048576,
  "tool_calling": true,
  "open_source": true,
  // ── Pricing ──────────────────────────────────────
  "direction": "Input",
  "original_price": "$0.27 / 1M tokens",
  "normalized_price": 0.00027,
  "normalized_unit": "per 1,000 tokens",
  "billing_method": "Token-based",
  // ── Delivery & Geography ─────────────────────────
  "delivery_method": "Real-Time",
  "region": "North America",
  "country": "United States",
  "verification_date": "2026-03-01",
  // ... 32 fields per SKU · 1,625 SKUs · 40 vendors
}

LIVE · 200 OK · 12ms

application/json · 1,625 SKUs · 40 vendors

32 fields per SKU — full schema access

sku_id vendor_name vendor_id vendor_type model_name model_id sku_plan_name modality modality_subtype model_family context_window tool_calling open_source direction billing_method original_price normalized_price normalized_unit normalization_formula normalization_method delivery_method deployment_method contracting_method region country vendor_url pricing_url source_url pricing_notes verification_date run_id null_reasons

What's In The Feed

Comprehensive Coverage, Every Week

—

SKUs Indexed

Input, cached input, and output pricing for every model

—

Models Tracked

Across text, image, audio, video, voice, and multimodal

—

Vendors Covered

From hyperscalers to specialized inference providers

—

Modalities

Text, Image, Audio, Video, Voice, and Multimodal

—

Regions

North America, Europe, Greater China, Advanced Asia, and Middle East

—

Countries

Country-level pricing granularity where vendors differentiate

Full Schema

32 Fields, Normalized and Queryable

Identity

sku_id

Unique identifier for each pricing record

vendor_id

Internal vendor identifier

vendor_name

Provider offering the model

vendor_type

Model Developer, Inference Platform, or Cloud Marketplace

model_id

Vendor's canonical model identifier

model_name

Standardized model name across vendors

sku_plan_name

Vendor's plan or product name for the SKU

Model Intelligence

modality

Text, Image, Audio, Video, Voice, or Multimodal

modality_subtype

Specific function within modality (e.g., Chat Completion, Embedding)

model_family

Model family grouping (e.g., Llama 4, GPT-4, Claude 3)

context_window

Maximum context window in tokens

tool_calling

Boolean — whether the model supports function/tool calling

open_source

Boolean — whether the model weights are publicly available

Pricing

direction

Input, Output, or Cached Input pricing

billing_method

Token-based, character-based, per-image, per-second, etc.

original_price

Price as listed by the vendor in original format

normalized_price

Price converted to standardized comparable unit

normalized_unit

Standardized unit (e.g., per 1,000 tokens)

normalization_formula

Exact calculation used for price conversion

normalization_method

Method applied (e.g., Standard Token, Converted-to-Optimal Unit)

Delivery & Access

delivery_method

Real-Time, Batch, or Provisioned

deployment_method

Cloud API (Shared), Dedicated, On-Premise, etc.

contracting_method

Self-serve, Sales-assisted, or Enterprise agreement

Geography

region

Geographic region (North America, Europe, Greater China, etc.)

country

Country-level granularity where vendors differentiate pricing

Provenance & Verification

vendor_url

Vendor's main website

pricing_url

Direct link to vendor's pricing page

source_url

Link to model documentation or specifications

pricing_notes

Additional context or extraction notes

verification_date

Date the pricing was last verified from source

run_id

Unique identifier for the indexing run

null_reasons

JSONB column explaining why a field is null: not_applicable, not_disclosed, or not_available

Integration

How It Works

Authenticate

Receive your API key after onboarding

Query

Filter by model, vendor, region, modality, or any of 32 fields

Receive

Get structured JSON with 32 normalized fields per SKU

Integrate

Pipe into models, dashboards, or routing systems

Who It's For

Built for Systems That Need Pricing Data

Quant & Research Teams

Feed inference pricing signals into quantitative models. Track cost compression trends. Build alternative data strategies around AI infrastructure economics.

Infrastructure & Platform Teams

Optimize inference routing with live cost data. Benchmark vendor pricing across regions. Automate procurement decisions with structured feeds.

AI-Native Companies

Monitor competitor pricing in real time. Track market dynamics across modalities. Inform your own pricing strategy with comprehensive market intelligence.

Get Started

Ready to Connect?

Schedule a walkthrough to see the full dataset, discuss your use case, and explore integration options.

SECURITY & COMPLIANCE: ATOM is built on SOC 2 Type II certified platforms and ISO 27001:2022 certified data storage infrastructure. Our technology stack undergoes regular third-party security audits by accredited certification bodies. Our infrastructure providers maintain compliance with SOC 2, ISO 27001, ISO 27017 (Cloud Security), ISO 27701 (Privacy), and GDPR. ATOM inherits these security controls and data protection practices through our platform architecture.