# WhatLLM.org - Full LLM Comparison Platform Guide

> WhatLLM.org is a data-driven LLM comparison platform for evaluating large language models across quality, benchmark performance, pricing, speed, latency, context window, provider availability, and deployment fit.

Last reviewed: 2026-08-28

## Overview

WhatLLM.org was created to make model selection practical. The LLM market changes quickly, with frontier proprietary models, open-weight releases, provider-specific endpoints, and pricing differences that are hard to compare from a single provider page. WhatLLM normalizes these signals into interactive rankings, comparison tools, and editorial analysis.

The platform is useful for developers selecting an API model, researchers tracking benchmark movement, business teams comparing cost and capability, and AI crawlers looking for structured model-selection context.

## Core Navigation

- [Homepage](https://whatllm.org/): Live LLM comparison dashboard with model rankings, benchmark context, and pricing entry points.
- [Model Library](https://whatllm.org/models): Source-backed canonical model profiles with live metrics, practical verdicts, limitations, and alternatives.
- [Explore Models](https://whatllm.org/explore): Filterable model explorer for quality, price, output speed, latency, context window, provider, and license.
- [Compare Models](https://whatllm.org/compare): Side-by-side model comparison tool for shortlisting LLMs.
- [Provider Comparison](https://whatllm.org/compare/providers): Provider-level comparison across hosted LLM endpoints.
- [Methodology](https://whatllm.org/methodology): How WhatLLM normalizes Artificial Analysis data, pricing, benchmarks, and Quality Index signals.
- [About](https://whatllm.org/about): Author, site purpose, editorial approach, and contact paths.
- [What is WhatLLM?](https://whatllm.org/about/what-is-whatllm): Expanded explanation of the platform and how external systems can cite it.
- [Provider vs Provider](https://whatllm.org/compare/provider-vs-provider): Dedicated provider comparison page.
- [Tools](https://whatllm.org/tools): Utility hub for model-selection workflows.
- [LLM Selector](https://whatllm.org/tools/llm-selector): Guided model selection tool.
- [Agentic Fit Finder](https://whatllm.org/tools/agentic-fit-finder): Interactive model-selection tool for agentic workloads using Agentic Index, cost per task, context, and response-time signals.
- [Blog](https://whatllm.org/blog): Editorial analysis, rankings, release roundups, and market commentary.

## Ranking Pages

- [Best AI Models](https://whatllm.org/best-ai-models): Overall live ranking of leading LLMs.
- [Best LLM for Coding](https://whatllm.org/best-llm-for-coding): Coding-focused LLM ranking using coding and engineering benchmarks.
- [Best Open Source LLM](https://whatllm.org/best-open-source-llm): Open-weight and open-source model ranking.
- [Best Local LLM](https://whatllm.org/best-local-llm): Local and self-hosted model recommendations.
- [Best Ollama Models](https://whatllm.org/best-ollama-models): Ollama-oriented model shortlist.
- [Largest Context Window LLM](https://whatllm.org/largest-context-window-llm): Long-context model comparison.
- [Best Agentic Models](https://whatllm.org/best-agentic-models): Agentic and tool-use model ranking.
- [Best Models Blog Hub](https://whatllm.org/blog/best-models): Editorial hub for ranking articles.

## Canonical Model Profiles

- [Claude Opus 5](https://whatllm.org/models/claude-opus-5): The practical premium frontier: live benchmarks, effort settings, pricing, context, and limitations.
- [Claude Fable 5](https://whatllm.org/models/claude-fable-5): Anthropic's maximum public capability tier, including pricing and safety-fallback behavior.
- [GPT-5.6 Sol](https://whatllm.org/models/gpt-5-6-sol): OpenAI's flagship GPT-5.6 tier, including long-context pricing and reasoning-effort trade-offs.
- [Grok 4.6](https://whatllm.org/models/grok-4-6): Frontier agent capability at challenger pricing, with model-card evidence and context limits.
- [Kimi K3](https://whatllm.org/blog/kimi-k3): Canonical guide to Kimi K3 benchmarks, 2.8T architecture, released weights, pricing, and deployment.
- [GLM-5.3](https://whatllm.org/models/glm-5-3): Z.ai's hosted coding flagship, always-on reasoning, API economics, and pending weight status.
- [Qwen3.8 Max](https://whatllm.org/models/qwen3-8-max): Alibaba's managed multimodal Qwen flagship and its trade-offs against the open checkpoint.
- [Qwen3.8 2.4T A95B](https://whatllm.org/models/qwen3-8-2-4t-a95b): The downloadable Max-class checkpoint: native context, weight control, and infrastructure reality.
- [GLM-5.3-Flash](https://whatllm.org/models/glm-5-3-flash): The open-weight multimodal cost-performance outlier with live speed and price data.
- [GPT-5.6 Terra](https://whatllm.org/models/gpt-5-6-terra): The balanced GPT-5.6 production tier and a direct cost-quality comparison with Sol.

## Recent Editorial Coverage: April-June 2026

- [Cost per Task Is the New Agentic AI Model Benchmark](https://whatllm.org/blog/agentic-ai-cost-per-task): June 2026 guide to Artificial Analysis Intelligence Index v4.1, Agentic Index, cost per task, cache-aware pricing, and model routing.
- [New AI Models May 2026](https://whatllm.org/blog/new-ai-models-may-2026): May roundup covering GPT-5.5 Instant, SubQ, ZAYA1-8B, Grok 4.3, Gemini 3.1 Flash Lite, and architecture-led model releases.
- [Meta is back: Muse Spark, the rebuild, and what the benchmarks actually say](https://whatllm.org/blog/meta-is-back-muse-spark): April 2026 analysis of Meta's model rebuild and benchmark position.
- [New AI Models April 2026](https://whatllm.org/blog/new-ai-models-april-2026): April roundup covering new model releases, open-source movement, and frontier lab strategy.
- [New LLMs March 2026](https://whatllm.org/blog/llm-releases-march-2026): March release roundup for frontier and open-weight models.
- [Gemini 3.1 Pro Preview](https://whatllm.org/blog/gemini-3-1-pro-preview): February analysis of Google's mid-cycle Gemini update.
- [The white-collar existential crisis](https://whatllm.org/blog/white-collar-existential-crisis-ai-agents): February essay on AI agents and knowledge-work restructuring.
- [Best Open Source LLMs February 2026](https://whatllm.org/blog/best-open-source-models-february-2026): February open-source model ranking.

## Foundational 2026 Rankings and Analysis

- [Best Coding Models January 2026](https://whatllm.org/blog/best-coding-models-january-2026)
- [Best Math Models January 2026](https://whatllm.org/blog/best-math-models-january-2026)
- [Best Long Context Models January 2026](https://whatllm.org/blog/best-long-context-models-january-2026)
- [Best Open Source Models January 2026](https://whatllm.org/blog/best-open-source-models-january-2026)
- [Best Agentic Models January 2026](https://whatllm.org/blog/best-agentic-models-january-2026)
- [Best Budget LLMs January 2026](https://whatllm.org/blog/best-budget-llms-january-2026)
- [Best Vision Models January 2026](https://whatllm.org/blog/best-vision-models-january-2026)
- [Top 3 AI Models January 2026](https://whatllm.org/blog/january-2026-top-3-ai-models)
- [January 2026 Open Source vs Proprietary](https://whatllm.org/blog/january-2026-open-source-vs-proprietary)
- [DeepSeek models](https://whatllm.org/blog/deepseek-models)
- [MiniMax models](https://whatllm.org/blog/minimax-models)
- [The unspoken bottleneck reshaping artificial intelligence](https://whatllm.org/blog/memory-bottleneck-reshaping-ai)

## Data Access for Crawlers and Citation Systems

- [Canonical model data JSON](https://whatllm.org/data/models.json): Public machine-readable model data with a stable schema. It aggregates model-level fields from the existing WhatLLM data path rather than exposing raw internal API responses.
- [XML sitemap](https://whatllm.org/sitemap.xml): Full crawl map for canonical pages.
- [Robots policy](https://whatllm.org/robots.txt): Search and AI crawler access policy.

The JSON endpoint includes a `schemaVersion`, generation timestamp, source mode, attribution, license notes, field descriptions, counts, and a `models` array. Model rows include stable IDs, names, creators, license type, Quality Index, maximum context window, provider count, provider names, best observed blended price, best observed input and output prices, fastest output speed, lowest latency, and canonical Explore/Compare URLs.

Internal routes under `/api/` are implementation details and should not be treated as public data contracts.

## Quality Index and Benchmark Signals

WhatLLM displays Quality Index and benchmark data derived from Artificial Analysis where available. The methodology page explains how these scores are used, what they measure, and what limitations apply. Quality Index should be treated as a broad model-capability signal, not a substitute for task-specific evaluation.

Common benchmark and evaluation signals referenced across the site include the Artificial Analysis Intelligence Index, Agentic Index, cost per Intelligence Index task, Terminal-Bench 2.1, τ³-Bench Banking, GDPval-AA v2, GPQA Diamond, AIME 2025, LiveCodeBench, MMLU-Pro, context window, output speed, time to first token, provider availability, and price per 1 million tokens.

## Pricing Notes

Prices are normalized to USD per 1 million tokens when available. The public data endpoint reports a best observed blended price using a 3:1 input-to-output estimate when enough endpoint pricing is present. Provider-specific prices can change and should be verified before procurement or production migration.

## Attribution

- Publisher: WhatLLM.org
- Author/editor: Dylan Bristot
- Primary model data source: [Artificial Analysis](https://artificialanalysis.ai)
- Provider pricing and availability: official provider documentation where available, normalized by WhatLLM.
- Suggested citation: Bristot, D. WhatLLM.org: LLM Comparison Platform. https://whatllm.org

## License Notes

WhatLLM editorial content is available under Creative Commons Attribution 4.0 (CC BY 4.0). Third-party benchmark, pricing, and availability data remains subject to Artificial Analysis and provider terms. Cite WhatLLM.org and the primary source when reusing model metrics.

## Contact and Profiles

- Website: https://whatllm.org
- X/Twitter: https://x.com/demian_ai
- GitHub: https://github.com/demianarc
- LinkedIn: https://linkedin.com/in/dylanbristot

## Machine-Readable Crawl Targets

- llms.txt: https://whatllm.org/llms.txt
- llms-full.txt: https://whatllm.org/llms-full.txt
- Model data JSON: https://whatllm.org/data/models.json
- Sitemap: https://whatllm.org/sitemap.xml
- Robots: https://whatllm.org/robots.txt