Curated. Verified. Transparent.

Models

Source-backed intelligence on today’s documented AI models.

Provider claims, technical limits and pricing are recorded with their evidence. Where a fact cannot be verified, the record says so.

Published records: 4
Providers: 4
Latest verification: August 23, 2026

Featured model dossier

Editorial selection

Anthropic

Claude Opus 5

Anthropic's model for complex agentic coding and enterprise work, documented with a 1M-token context window and 128k maximum output tokens on the synchronous Messages API.

Open model dossier

Status: active
Context window: 1,000,000 tokens
Inputs: Text and Image
Last verified: August 19, 2026
Sources: 2 cited

Browse by provider

4 documented

Anthropic1 model record→
Google1 model record→
OpenAI1 model record→
xAI1 model record→

Model index

4 verified records

A compact comparison of facts documented by primary provider sources. Open a record for pricing, field-level disclosure states and complete provenance.

AI model reference records
Model	Provider	Status	Context	Inputs	Last verified	Open record
Claude Opus 5`claude-opus-5`	Anthropic	active	1,000,000 tokens	Text and Image	August 19, 2026	→
Gemini 2.5 Pro`gemini-2-5-pro`	Google	active	1,048,576 tokens	Text, Image, Audio and Video	August 23, 2026	→
GPT-5.6 Sol`gpt-5-6-sol`	OpenAI	active	1,050,000 tokens	Text and Image	August 23, 2026	→
Grok 4.6`grok-4-6`	xAI	active	500,000 tokens	Text	August 23, 2026	→

Evidence standard

How to read a model record.

Absence and uncertainty are part of the record, not gaps to disguise.

Known: Published by a source cited on the model page.
Unknown: Not yet verified; no plausible value is substituted.
Not disclosed: The provider does not publish the value.
Estimated: Derived from evidence and accompanied by its method.