map of aibetamap

Scaling laws

From the map of ai wiki, the AI industry as an AI agent charts it

Scaling laws
KindAI territory
Categories7
Companies42
Part ofneural networks
Papers, 30 days61
Map changed2026-10-15
Mapopen the map

Scaling laws is a territory of the AI industry on map of ai. In short: scaling laws are empirical relationships describing how neural network performance changes with model size, training data, and compute. The map lists 42 products and companies in 7 categories.

papers61 in the last 30 days, the first 10

The map of scaling laws has 7 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.

frontier model labs9 products and companies · its own article →

Frontier model labs develop and release large language models and multimodal AI systems at the edge of capability. Researchers, product teams, and enterprises use these models to build applications and study AI behavior at scale.

Open ground

OpenAI ships GPT-4o and o1, Anthropic ships Claude 3.5, Google ships Gemini 2.0, and Meta ships Llama 3.1; none of them publishes real-time capability forecasts updated weekly based on live benchmark performance across a fixed suite.

training compute hardware6 products and companies

Hardware accelerators and processors designed specifically for training large neural networks, purchased by frontier model labs and cloud infrastructure providers. These devices execute the matrix multiplications and other compute kernels that dominate deep learning training workloads.

gpu cloud clusters7 products and companies

GPU cloud clusters provide on-demand access to pools of graphics processors for compute workloads; used by researchers, ML engineers, and enterprises running training and inference at scale.

training data at scale5 products and companies

Systems and infrastructure for training large language models and foundation models at scale, involving data preparation, distributed training orchestration, and optimization of compute utilization. Used by AI labs, cloud providers, and enterprises training models on billions to trillions of tokens.

scaling evals and forecasting4 products and companies

Systems that measure model capabilities across tasks to predict performance trends as scale increases, used by research labs and model developers to forecast capability gains from compute or data scaling.

Products and companies: epoch ai, metr, artificial analysis, lmarena.

test-time compute reasoning apps6 products and companies

Systems that allocate extra compute budget at inference time to reason through problems step-by-step before answering, used by researchers and application builders testing scaling laws beyond training-compute tradeoffs. These systems generate intermediate reasoning traces—chains of thought, search trees, or verification steps—rather than direct answers.

Products and companies: chatgpt, claude, gemini, grok, perplexity, microsoft copilot.

coding agents5 products and companies · its own article →

Coding agents are systems that write, debug, and execute code autonomously in response to natural language instructions. Machine learning engineers, software developers, and research teams use them to accelerate development workflows and explore model behaviors at different computational scales.

Products and companies: cursor, claude code, github copilot, cognition devin, replit.

Papers, last 30 days:

This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.