AI Datasets & Benchmarks
Free, structured datasets for AI decision-makers. Cost benchmarks, model comparisons, and performance data. Updated quarterly by Salt Technologies AI engineers.
AI Development Cost Benchmark 2026
Average cost ranges for AI development projects in 2026, broken down by project type, complexity level, timeline, and team size. Based on market analysis and delivery data from Salt Technologies AI.
LLM Model Comparison for Enterprise Use Cases (2026)
Structured dataset comparing 16 large language models across 22 fields: pricing per million tokens, benchmark scores (MMLU, HumanEval, MATH, MT-Bench), latency, context windows, API features, and recommended enterprise use cases. Covers 7 providers: OpenAI, Anthropic, Google, Meta, Mistral AI, DeepSeek, and Cohere.
Vector Database Performance Benchmark 2026
Structured dataset comparing 10 vector databases across 19 fields: query latency (p50/p99 in milliseconds), indexing throughput, maximum dimensions, index algorithms, distance metrics, hybrid search support, multi-tenancy, ACID compliance, SDK support, and pricing. Covers Pinecone, Qdrant, Weaviate, Milvus, pgvector, ChromaDB, Elasticsearch, MongoDB Atlas, Redis, and Supabase.
Fortune 500 AI Adoption Tracker 2026
Structured dataset tracking AI adoption across 50 Fortune 500 companies in 2026. Covers AI investment levels, maturity scores, use cases, AI leadership, patents, partnerships, and public announcements. Spans 14 industries including technology, financial services, healthcare, retail, and energy.
Fortune 500 Technology Stack Directory 2026
Structured dataset mapping the technology stacks of 100 Fortune 500 companies across cloud providers, enterprise platforms, programming languages, databases, and AI/ML tools. Covers the top 100 by revenue plus select high-profile entries. 12 fields per record.
Fortune 500 Companies Directory 2025: Complete List
Complete directory of all 500 Fortune 500 companies ranked by revenue for fiscal year 2024. Includes rank, company name, industry, revenue (USD millions), employee count, headquarters city, and state. 7 fields per record, 500 records.
Why We Publish This Data
CTOs evaluating AI investments deserve transparent, structured data, not vague blog posts with "it depends" answers.
Based on real delivery data
Numbers come from 800+ projects delivered by Salt Technologies, not surveys or estimates.
Updated quarterly
AI pricing and performance change fast. We re-benchmark every quarter so the data stays current.
Free to use and cite
Published under CC BY 4.0. Use the data in your own research, presentations, and proposals with attribution.
Methodology published
Every dataset includes a full methodology section so you can evaluate the data quality yourself.
Need AI expertise, not just data?
Start with a $3,000 AI Readiness Audit. Get a clear roadmap in 1-2 weeks.