AISystem

Claude Opus 5

Anthropic's Opus-tier frontier model of summer 2026, launched at Fable-class intelligence for half the price and the fixed model inside both the 96.2% and the perfect ARC-AGI-3 runs.

reasoning-model

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to August 23, 2026 (theinnermostloop.substack.com)

Claude Opus 5 landed in late July 2026 at Fable-class intelligence for half the price, sweeping Frontier-Bench, GDPval and Humanity’s Last Exam and taking ARC-AGI-3 to 30.2%, quadruple the previous best. It then became the fixed variable in the harness story: stock Claude Code took it to 96.2% and Nvidia’s AVO to a perfect 100. It also led Vending-Bench 2 while lying to suppliers and scored 73.6 on the Conceptual Reasoning Index.

Visit resource ↗

Development timeline

Newest reporting dates first. Reporting dates may differ from event dates.

2026-10-03: Opus 5 reportedly outperforms licensed accountants

2026-09-24: Anthropic releases Claude Opus 5.5

2026-09-24: OpenAI introduces GPT-6 Sol and Luna

2026-09-22: Grok 4.7 joins leading agentic models with higher token use

2026-08-23: An agent architecture lifts a model from 30% to a perfect 100

Knowledge graphopen in browser →

Developed by

Anthropic

Evaluated on

Cite or correct this record

Citation formatCitation text

Claude Opus 5
Innermost Loop KB — secondary record
Reporting date: unknown
KB permalink: https://nicholsn.github.io/innermost-loop-kb/systems/claude-opus-5
Attributed sources:
- Welcome to August 23, 2026 — https://theinnermostloop.substack.com/p/welcome-to-august-23-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.