sourc.dev — AI and developer data observatory
Data observatory · AI infrastructure · Updated daily
Set your needs.
The data decides.
The data decides.
LLMs, APIs, tools, and AI infrastructure — tracked, measured, and compared over time. Sourced. Verified. No opinions.
Compare & choose
I need to find the right model now
Filtering, pricing, EU residency — all in one sortable table.
Go to LLM table
Market intelligence
I want to understand how the market moves
Leaderboards, historical pricing, ecosystem scores.
See leaderboards
Monitor changes
I want to know when something changes
Price drops, new models, drift flags. Subscribe free.
Follow change feed
Source-verified every 24h
·
Historical archive from 2026
·
No opinion expressed
LLMs tracked
30
verified 2026-03-27
Relationships mapped
933
verified · sourced · growing
Price reduction 2020–25
97%
$60 → $0.10 per 1M tokens
Cheapest input today
$0.07
Gemini 1.5 Flash · per 1M tokens
Context window growth
244×
4,096 → 1,000,000 tokens
The network
4 properties
view all → llms
Language models — pricing, context windows, benchmarks, drift scores
30
live
tools
Developer tools — integrations, pricing tiers, EU residency, categories
129
live
apis
AI APIs — inference providers, model routing, pricing, and documentation
37
live
saas
SaaS platforms — vector databases, observability, cloud infrastructure
69
live
Latest changes
feed →
Monitor the feed
Alerts when any model you depend on changes price, context, or drift score.
Free member access
LLM pricing 2020–2025 · verified 2026-03-01
$60 → $0.10
One million tokens. The steepest price collapse in software infrastructure history.
244×
Context windows since 2020
GPT-3 launched with 4,096 tokens. Gemini 1.5 Pro runs to one million. The same document that needed 244 API calls now needs one.
Verified 2026-03-01 · sourc.dev pipeline
1668
Riksbanken interest rate data
The world's longest continuous economic time series — available via free API. Some infrastructure is older than most countries.
Verified 2026-02-14 · riksbank.se
65%
Organisations using gen AI 2024
Up from 33% twelve months prior. The fastest technology adoption curve McKinsey has documented in their 30-year history.
Verified 2025-11-01 · McKinsey Global Survey
Today's most-watched
auto-ranked by views
live 2026-03-27 LLM
Gemini 1.5 Flash
★ best_value ↓ cheapest
Input /1M
$0.07
Output /1M
$0.30
Context
1.0M
Quality
84
Leaderboards
auto-computed daily
all leaderboards → ◈ Cost guide →
Building at volume? These 3 models optimise for cost at your scale.
Input your monthly token estimate and get a data-driven cost comparison across all 30 models.
🇪🇺 EU guide →
Operating under GDPR or EU AI Act compliance?
18 models offer verified EU data residency. Get the compliance guide.
Editorial · Deep dives
Price history
The price collapse
$60 → $0.10 · 97% reduction · 2020–2025
LLM prices have collapsed faster than anything in software infrastructure history. Here is the data, the timelines, and why it isn't over yet.
Read the story
Context windows
244× expansion in four years
4,096 → 1,000,000 tokens
From one document per call to an entire archive. The history behind the context window's growth — and what it actually means for what you can build.
Read the story
Model generations
GPT · Claude · Gemini · Llama
Full family timelines
Every family, every generation, every price movement and capability jump — chronological, source-linked, without editorial judgment.
Read the story