Model Class | LLMS3

Embedding Model Model Class

A class of model that converts unstructured data (text, images, audio) into fixed-dimensional vector representations suitable for …

7 3

General-Purpose LLM Model Class

A large language model for broad text tasks. In scope when applied to metadata extraction, summarization, schema inference, or que…

10 3

Code-Focused LLM Model Class

An LLM specialized for code understanding and generation. A subtype of General-Purpose LLM with enhanced ability to work with stru…

6 3

Small / Distilled Model Model Class

A compact model (typically under 10B parameters) suitable for local or edge deployment, often distilled from a larger model to ret…

2 2

Reranker Models Model Class

A class of model that re-scores and re-orders retrieval results from vector search, improving precision by applying a more expensi…

4 2

Metadata Extraction Models Model Class

Specialized models for extracting structured metadata (entities, dates, categories, relationships) from unstructured documents sto…

3 2

Document Parsing / OCR / VLM Models Model Class

Models that convert scanned documents, images, and PDFs stored in S3 into structured, machine-readable text. Includes OCR engines,…

3 3

Anomaly Detection Models Model Class

Models that identify unusual patterns in S3 access logs, storage metrics, API call patterns, and billing data — flagging potential…

5 2

Data Quality Validation Models Model Class

Models that assess the quality, completeness, and consistency of data arriving in S3 — checking for missing values, format violati…

4 2

Classification / Tagging Models Model Class

Models that automatically categorize S3 objects by content type, sensitivity level, domain, or business unit — enabling automated …

5 2

Cost Optimization Models Model Class

Models that analyze S3 usage patterns — access frequency, storage class distribution, request types, egress volumes — and recommen…

7 2

Policy Recommendation Models Model Class

Models that analyze existing IAM policies, bucket policies, and access patterns for S3 environments, recommending improvements for…

5 2

Mixture-of-Experts (MoE) Model Class

A neural-network architecture pattern where each input token is dynamically routed to a small subset of specialized "expert" sub-n…

5

Kimi K2 Model Class

Frontier open-weight Mixture-of-Experts large language model from Moonshot AI. Architecture: **1T total parameters, 32B activated …

8

DeepSeek V3 Model Class

Open-weight 671B-parameter Mixture-of-Experts language model from DeepSeek AI. **37B activated per token** (5.5% activation ratio)…

15

DeepSeek-R1 Model Class

Reasoning-focused open-source language model built on the DeepSeek-V3 base. Inherits the 671B total / 37B active MoE architecture …

3

GLM-5 Model Class

Open-weight frontier MoE model from Zhipu AI (清华系 Beijing-based AI lab). **744B total parameters, 40-44B active per inference toke…

4

Llama 4 Model Class

Meta's open-weight LLM family, released April 5, 2025 — the first Llama models to use Mixture-of-Experts (MoE) architecture and th…

4

Qwen3 Model Class

Alibaba's open-weight large language model family launched April 2025 and evolved through multiple 2026 releases. Range covers six…

4

DeepSeek V4 Model Class

DeepSeek's flagship V3 successor, served as V4-Pro (1M context) and V4-Flash, whose 75% price cut became the permanent list price …

9 3

Claude Fable 5 Model Class

Anthropic's June 9, 2026 frontier model — Fable 5 (GA, `claude-fable-5`, $10/$50 per MTok) and Mythos 5 (same model, safeguards li…

4 1