All terms · A-Z

model_provider · valuation $NaNB · international.

Vertically stacked NAND flash · used for AI training data storage · 300+ layers in 2026 flagship nodes.

NVIDIA's ? TFLOPS AI accelerator with ? GB HBM.

An AI system that plans, uses tools, and takes multi-step actions to accomplish goals · not just a chat turn.

The research field focused on making AI systems pursue intended goals safely and reliably.

BenchGecko's 0-1000% composite score measuring AI sector valuation vs fundamentals.

The capital expenditure AI labs and hyperscalers spend on chips, datacenters, and training clusters.

model_provider · valuation $NaNB · international.

AI researcher.

Paul Gauthier's terminal-based git-aware AI pair programmer · commits with context of the whole repo.

Aider polyglot is a coding benchmark tracked by BenchGecko across every frontier and open-weight model.

model_provider · valuation $NaNB · international.

Attention with Linear Biases · position encoding via attention-score penalties instead of positional embeddings.

public_big_tech · valuation undisclosed · international.

public_big_tech · valuation undisclosed · international.

infrastructure · valuation $NaNB · international.

AMD's next-gen data-center GPU · late 2025 announcement · targets H200/Blackwell competition with HBM4.

AI researcher.

ANLI is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

model_provider · valuation $NaNB · international.

infrastructure · valuation $NaNB · international.

developer_tools · valuation $NaNB · international.

APEX-Agents is a agentic benchmark tracked by BenchGecko across every frontier and open-weight model.

infrastructure · valuation $NaNB · international.

Running the same open-weight model on multiple providers to exploit price differences · up to 30× savings.

ARC AI2 is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

ARC-AGI is a reasoning benchmark tracked by BenchGecko across every frontier and open-weight model.

ARC-AGI-2 is a reasoning benchmark tracked by BenchGecko across every frontier and open-weight model.

Annualized Recurring Revenue · the standardized metric SaaS and AI companies report to investors.

AI researcher.

Huawei's ? TFLOPS AI accelerator with ? GB HBM.

infrastructure · valuation $NaNB · international.

AutoGen is Microsoft's open-source framework for building multi-agent AI applications.

AutoGPT is the open-source autonomous agent that kicked off the entire category.

NVIDIA's ? TFLOPS AI accelerator with ? GB HBM.

NVIDIA's ? TFLOPS AI accelerator with ? GB HBM.

model_provider · valuation $NaNB · international.

Balrog is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

infrastructure · valuation $NaNB · international.

Submitting many prompts at once in a single batch job · providers discount 50% on delayed batch completion.

BBH is a reasoning benchmark tracked by BenchGecko across every frontier and open-weight model.

infrastructure · valuation $NaNB · international.

Net cash burned divided by net new ARR · < 1× is healthy · AI startups often run 2-5× pre-scale.

Monthly cash outflow · the speed at which an AI company spends investor capital before hitting breakeven.

Bring Your Own Key · you pay the model provider directly, the app just routes your requests.

The fraction of input tokens served from provider-side prompt cache · directly impacts effective pricing.

Discounted input tokens for prompts the provider has already processed · 50-90% off on repeat prefix.

CadEval is a coding benchmark tracked by BenchGecko across every frontier and open-weight model.

infrastructure · valuation $NaNB · international.

A prompting technique (and trained behavior) where the model shows step-by-step reasoning before the final answer.

application · valuation $NaNB · international.

Crowdsourced head-to-head AI model comparison · humans vote on anonymous outputs and Elo ratings rank the models.

Chess Puzzles is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

Anthropic's foundation-model family · Opus / Sonnet / Haiku tiers plus Extended Thinking variants.

Claude Code is a command-line coding agent built by Anthropic that lives in your terminal and edits files, runs commands, and searches your codebase with direct Claude model access.

AI researcher.

Open-source VSCode coding agent · forked as Roo Code and Continue, built on user-supplied API keys.

OpenAI's open-source terminal coding agent · runs GPT models against your repo from the shell.

Sourcegraph's enterprise coding AI · deep codebase context from graph-aware search.

model_provider · valuation $NaNB · international.

The maximum number of tokens a model can process in one request · input prompt plus generated output.

Continue is an open-source IDE extension for building your own AI code assistant with custom models, context providers, and slash commands.

Open-source IDE-native coding assistant · inline autocomplete + chat across VSCode and JetBrains.

application · valuation $NaNB · international.

infrastructure · valuation $NaNB · international.

TSMC's next-gen advanced packaging · enables 12+ HBM stacks and multi-die GPUs · used on B300, MI400.

TSMC's previous-gen advanced packaging · used on H100 / A100 · being phased out for CoWoS-L.

CrewAI is a framework for orchestrating autonomous AI agents with roles, goals, and tools.

infrastructure · valuation $NaNB · international.

Cerebras appliance AI system.

CSQA2 is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

Cursor is a fork of VS Code rebuilt around an AI pair programmer.

% of revenue from top customers · AI labs often 40-60% from top 10 vs SaaS median <20%.

Cybench is a coding benchmark tracked by BenchGecko across every frontier and open-weight model.

AI researcher.

The requirement that data be stored and processed in a specific geographic region.

infrastructure · valuation $NaNB · international.

? GB/s memory at ? GB per stack · from major vendors.

Next-gen DRAM standard · 8.8-17.6 Gbps · used in servers + CPUs feeding GPU clusters · JEDEC finalized 2025.

DeepResearch Bench is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

Hangzhou-based lab behind V3 and R1 · the MoE + RL-reasoning recipe that crashed open-weight prices.

AI researcher.

Cognition AI's autonomous software engineer agent · browses, writes, and deploys code end-to-end.

NVIDIA node AI system.

NVIDIA rack AI system.

NVIDIA rack AI system.

Reduction in existing shareholder ownership % after new equity issuance · typical 15-25% per round.

Training a small model to mimic a large one, preserving most quality at a fraction of the size.

application · valuation $NaNB · international.

AI researcher.

A dense vector representation of text, image, or audio in high-dimensional space · the backbone of RAG and search.

Employee Stock Ownership Plan · pool of shares set aside for employees · typical 10-20% of cap table.

EU regulation classifying AI systems by risk level and imposing compliance obligations on developers and deployers.

US federal government authorization for cloud and AI services handling agency data.

Fiction.

Further training a pre-trained model on domain data to adapt behavior without retraining from scratch.

infrastructure · valuation $NaNB · international.

The no-cost tier of an AI API · usually rate-limited, quota-capped, or branded-output-only.

FrontierMath-2025-02-28-Private is a math benchmark tracked by BenchGecko across every frontier and open-weight model.

FrontierMath-Tier-4-2025-07-01-Private is a math benchmark tracked by BenchGecko across every frontier and open-weight model.

Tool-use / function-call responses count as output tokens · structured JSON tool calls are billed normally.

The API mechanism letting models emit structured JSON to invoke external functions.

Intel's ? TFLOPS AI accelerator with ? GB HBM.

NVIDIA's ? TFLOPS AI accelerator with ? GB HBM.

NVIDIA's ? TFLOPS AI accelerator with ? GB HBM.

? GB/s memory at ? GB per stack · from major vendors.

Next-gen graphics DRAM · 32-40 Gbps · powers consumer + pro GPUs like RTX 5090, RTX 6000 Ada next-gen.

EU data protection regulation · imposes strict consent, residency, and rights requirements on any service processing EU data.

Google DeepMind's foundation-model family · multimodal from day one, long context leader.

GeoBench is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

AI researcher.

GitHub Copilot is the AI pair programmer from GitHub and Microsoft.

application · valuation $NaNB · international.

United States chip foundry · advanced nodes · $7.4B revenue.

Gross Merchandise Value · total dollar value of transactions · used for marketplace AI products.

model_provider · valuation $NaNB · international.

Block's open-source local coding agent · MCP-first, offline-capable, built by the company behind Square.

GPQA diamond is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

OpenAI's Generative Pre-trained Transformer lineage · GPT-1 through GPT-5 and onward.

GPT Engineer generates an entire codebase from a natural language spec.

infrastructure · valuation $NaNB · international.

Revenue minus inference + training + data costs · AI labs at 50-70%, vs pure SaaS at 80-90%.

Tethering AI responses to verified source material · RAG, citations, tool calls, retrieval.

GSM8K is a math benchmark tracked by BenchGecko across every frontier and open-weight model.

GSO-Bench is a coding benchmark tracked by BenchGecko across every frontier and open-weight model.

Runtime filters and policies enforced around model input and output · the safety layer outside the model itself.

AI researcher.

NVIDIA's ? TFLOPS AI accelerator with ? GB HBM.

NVIDIA's ? TFLOPS AI accelerator with ? GB HBM.

When an AI generates plausible-sounding but factually incorrect or fabricated content.

AI researcher.

application · valuation $NaNB · international.

? GB/s memory at ? GB per stack · from major vendors.

? GB/s memory at ? GB per stack · from major vendors.

? GB/s memory at ? GB per stack · from major vendors.

? GB/s memory at ? GB per stack · from major vendors.

HellaSwag is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

NVIDIA node AI system.

US law governing the handling of Protected Health Information (PHI) by healthcare providers and their vendors.

HLE is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

developer_tools · valuation $NaNB · international.

A 164-problem Python benchmark where the model writes a function from its docstring and passes unit tests.

Wafer-to-wafer bonding technique that skips bumps · enables 3D stacking for HBM4 and advanced packaging.

Implicit-convolution alternative to attention · linear-time long-range modeling · explored by Stanford + Together.

Running a trained model to generate outputs from new inputs · what happens every time you call an AI API.

AWS's third-gen inference chip · optimized for serving LLMs at scale with HBM3 and NeuronCore v3.

model_provider · valuation $NaNB · international.

Tokens you send to the model · priced separately from output tokens, usually cheaper.

infrastructure · valuation $NaNB · international.

United States chip foundry · advanced nodes · $0.9B revenue.

The network of capital flowing into AI · which funds + corporates dominate which tier of deals.

infrastructure · valuation $NaNB · international.

application · valuation $NaNB · international.

AI researcher.

Jules is Google's asynchronous coding agent.

Techniques to shrink the key-value cache during inference · sliding windows, quantization, eviction, sparsity.

developer_tools · valuation $NaNB · international.

LAMBADA is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

infrastructure · valuation $NaNB · international.

developer_tools · valuation $NaNB · international.

LangGraph is LangChain's framework for building stateful, multi-actor agents as graphs.

Time to first token · how fast the AI starts responding. Critical for interactive UX.

Lech Mazur Writing is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

A contamination-resistant benchmark that refreshes tasks monthly to prevent models from memorizing answers.

Meta's open-weight foundation-model family · the dominant open-source LLM lineage.

? GB/s memory at ? GB per stack · from major vendors.

Groq's ? TFLOPS AI accelerator with ? GB HBM.

Lifetime Value divided by Customer Acquisition Cost · >3× is healthy · AI consumer apps tracking 2-4×.

Microsoft's ? TFLOPS AI accelerator with ? GB HBM.

Microsoft rack AI system.

State-space sequence model · linear-time alternative to transformers · foundation of Mamba-2, Jamba, Zamba models.

Butterfly Effect Inc's autonomous general-purpose agent · went viral March 2025 for multi-step research and web automation.

infrastructure · valuation $NaNB · international.

MATH level 5 is a math benchmark tracked by BenchGecko across every frontier and open-weight model.

Model Context Protocol · an open standard for connecting AI models to external tools, data, and services.

model_provider · valuation $NaNB · international.

public_big_tech · valuation undisclosed · international.

MetaGPT models a software company with product managers, architects, engineers and QAs.

AMD's ? TFLOPS AI accelerator with ? GB HBM.

AMD's ? TFLOPS AI accelerator with ? GB HBM.

AMD node AI system.

AMD's ? TFLOPS AI accelerator with ? GB HBM.

infrastructure · valuation $NaNB · international.

public_big_tech · valuation undisclosed · international.

Microsoft's second-gen custom AI chip · successor to Maia 100 · targets Azure OpenAI inference at scale.

application · valuation $NaNB · international.

model_provider · valuation $NaNB · international.

Paris-based AI lab · open-weight (Mistral, Mixtral) and closed (Large) tiers, EU data sovereignty focus.

model_provider · valuation $NaNB · international.

A model architecture that routes each token to a subset of specialized experts, so only a fraction of parameters activate per forward pass.

MMLU is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

A harder version of MMLU with 10 answer choices, filtered noise, and more reasoning-heavy questions.

infrastructure · valuation $NaNB · international.

model_provider · valuation $NaNB · international.

Meta's ? TFLOPS AI accelerator with ? GB HBM.

Per-token pricing differs by input type · text tokens vs image tokens vs audio seconds vs video frames.

Models that process and generate multiple types of data · text, image, audio, video.

infrastructure · valuation $NaNB · international.

public_big_tech · valuation undisclosed · international.

model_provider · valuation $NaNB · international.

OpenBookQA is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

OpenHands (formerly OpenDevin) is a community-built autonomous software development agent that matches frontier closed-source agents on SWE-bench Verified.

OpenAI's computer-use agent · clicks, types, and browses websites on the user's behalf.

OSWorld is a agentic benchmark tracked by BenchGecko across every frontier and open-weight model.

OTIS Mock AIME 2024-2025 is a math benchmark tracked by BenchGecko across every frontier and open-weight model.

Tokens the model generates · priced 3-5× higher than input tokens due to sequential decoding.

Price-to-Sales ratio · valuation divided by annual revenue. The AI sector premium metric.

A flat fee per API call regardless of tokens · used for image generation, search, and some agent products.

application · valuation $NaNB · international.

developer_tools · valuation $NaNB · international.

PIQA is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

Plandex is an open-source terminal AI coding agent optimized for large, multi-file tasks.

PostTrainBench is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

The first and most expensive phase of LLM creation · training on trillions of tokens of internet text.

The practice of crafting model inputs to elicit better outputs · cheaper than fine-tuning, more reliable than hope.

infrastructure · valuation $NaNB · international.

Reducing the precision of model weights to shrink memory footprint and speed up inference.

Alibaba's open-weight model family · dominant in Chinese + multilingual benchmarks.

Retrieval-Augmented Generation · grounds model responses in external data retrieved at query time.

Anti-dilution provision that adjusts conversion price if a later round is at lower valuation · full vs weighted-average.

An LLM trained to spend extra compute at inference thinking before it answers, trading latency and cost for accuracy on hard tasks.

Reasoning models bill thinking tokens separately from output · can 2-10× effective cost per query.

model_provider · valuation $NaNB · international.

infrastructure · valuation $NaNB · international.

developer_tools · valuation $NaNB · international.

Replit's in-browser AI coding agent · turns a prompt into a deployed app without leaving the browser.

Pre-purchased dedicated throughput · flat hourly fee for guaranteed tokens-per-second from a specific model.

Annual revenue divided by headcount · leverage metric for AI labs where the ratio hits $3M+ vs SaaS at $300K.

VSCode coding agent forked from Cline · adds auto-mode, multi-tab orchestration, and Orchestrator agent pattern.

application · valuation $NaNB · international.

AI researcher.

infrastructure · valuation $NaNB · international.

infrastructure · valuation $NaNB · international.

South Korea chip foundry · advanced nodes · $12.5B revenue.

developer_tools · valuation $NaNB · international.

The empirical relationship between compute, data, parameters, and model capability.

ScienceQA is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

AI researcher.

SimpleBench is a reasoning benchmark tracked by BenchGecko across every frontier and open-weight model.

SimpleQA Verified is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

infrastructure · valuation $NaNB · international.

Attention pattern where each token attends only to a local window · used in Mistral, Gemma for long-context efficiency.

China chip foundry · advanced nodes · $7.8B revenue.

Smolagents is Hugging Face's minimalist agent library.

developer_tools · valuation $NaNB · international.

An audit framework certifying that a provider handles customer data with defined security controls.

Preemptible GPU capacity at deep discount · AWS Spot, GCP Spot, Azure Spot · 60-90% off on-demand but can be evicted.

model_provider · valuation $NaNB · international.

model_provider · valuation $NaNB · international.

A delegated sub-task agent spawned by a parent orchestrator · each runs in isolated context with specialized tools.

SWE-agent is the research agent from the Princeton NLP group behind SWE-bench.

The real-world coding benchmark · AI resolves actual GitHub issues in open-source Python repos.

SWE-Bench verified is a coding benchmark tracked by BenchGecko across every frontier and open-weight model.

Sweep is an AI junior developer that turns bug reports and feature requests into code changes as pull requests, running entirely from GitHub issues.

AI researcher.

application · valuation $NaNB · international.

Self-hosted AI coding assistant · Copilot-style autocomplete you can run on your own GPU.

An inference parameter controlling output randomness · 0 = deterministic, higher = more varied.

Terminal Bench is a coding benchmark tracked by BenchGecko across every frontier and open-weight model.

The Agent Company is a agentic benchmark tracked by BenchGecko across every frontier and open-weight model.

Total tokens per second a serving cluster can handle across all concurrent requests.

Volume discounts at defined usage thresholds · $X/M tokens below 1B tokens, $Y/M above.

infrastructure · valuation $NaNB · international.

The algorithm that splits text into tokens before a model reads it · BPE, SentencePiece, Tiktoken.

The fundamental units language models read and generate · roughly 3/4 of an English word per token.

A model invoking external functions or APIs during generation · the foundation of AI agents.

Google's ? TFLOPS AI accelerator with ? GB HBM.

Google pod AI system.

Google's ? TFLOPS AI accelerator with ? GB HBM.

Google pod AI system.

Google's ? TFLOPS AI accelerator with ? GB HBM.

AWS's third-gen custom training chip · 2× perf/watt vs Trainium 2 · powering Anthropic training clusters on AWS.

AWS's ? TFLOPS AI accelerator with ? GB HBM.

The neural network architecture (Vaswani et al., 2017) behind every modern large language model.

TriviaQA is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

AWS node AI system.

infrastructure · valuation $NaNB · international.

VideoMME is a multimodal benchmark tracked by BenchGecko across every frontier and open-weight model.

Pre-negotiated flat rate applied to all consumption · typical enterprise AI contract shape.

VPCT is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

developer_tools · valuation $NaNB · international.

developer_tools · valuation $NaNB · international.

WeirdML is a coding benchmark tracked by BenchGecko across every frontier and open-weight model.

Codeium's VSCode-fork IDE with "Cascade" multi-file agent · acquired by Cognition in 2025.

Windsurf is Codeium's purpose-built agentic IDE with Cascade, a flow-state collaborative agent that keeps context across your whole project.

Winogrande is a knowledge benchmark tracked by BenchGecko across every frontier and open-weight model.

Cerebras's ? TFLOPS AI accelerator with ? GB HBM.

model_provider · valuation $NaNB · international.

model_provider · valuation $NaNB · international.

AI researcher.

Zed is a Rust-built code editor from the Atom team.

Zed editor's built-in AI assistant · Rust-native, collaborative, MCP-first.

model_provider · valuation $NaNB · international.