LLM landscape

LLM Insights

Data-grounded views of the model landscape, refreshed automatically on a schedule. Each insight carries its own dataset, methodology, freshness status, and agent-written takeaways.

Models tracked

54

LLM cases

1,907

Tracked releases

385

AI labs

5

Updated 1 day ago

Highlights

Training compute vs capability

How published training compute relates to model capability, with measured rows, explicit assumptions, and a validation review queue.

54 models39 measured72 in reviewJul 28, 2026, 5:20 AM

Cost vs capability

What frontier capability costs to train — and which models deliver the most capability per dollar, traced by a Pareto efficiency frontier.

54 models priced39 measured cost30 open weightsJul 28, 2026, 5:20 AM

Capability per inference cost

Which models give the most capability per dollar to run — Epoch capability scores against published API prices per million tokens, with a cost-efficiency frontier.

51 priced models7 providersGPT-5.5 most capableJul 28, 2026, 5:20 AM

Benchmark performance

How models score on individual public evaluations — GPQA Diamond, Humanity's Last Exam, SciCode, AA-LCR and more — ranked as a per-benchmark bar chart.

10 evaluations1 populated13 modelsJul 28, 2026, 5:20 AM

Enterprise LLM adoption

Which LLM ecosystems show up in real, source-linked enterprise deployments — quarterly shares for Azure OpenAI, Bedrock, Gemini, Claude, and open weights.

Google Gemini & Vertex AI leads on discovery rate with 10 cases per 100 exploration runs in 2026-Q2.

1,907 LLM-classified cases15 quarters6 ecosystemsJul 28, 2026, 5:20 AM
OpenAIAnthropicGoogleAlibabaMicrosoft

Model release timeline

When and how often the large AI labs ship new models — release cadence per lab, kept current by the weekly catalogue refresh and the announcement-watcher agent.

385 tracked releases5 AI labsJul 24 latestJul 28, 2026, 5:20 AM