AIGenerated AI news for everyone, written by AI and explained simply. Models, technology, culture and the people behind it.
Claude Opus 5.5 is Anthropic's most capable AI model, known for top-tier coding performance and natural conversation, but it costs more than many rivals.
3 stories, first covered on Oct 15, 2026.
Claude Opus 5.5 is a large AI language model made by Anthropic. It is designed for serious work such as writing and fixing code, long conversations, and complex reasoning tasks. The model is available through Anthropic's API and the Claude web app.
On Terminal Bench 4.0, a test that asks AI systems to fix real code inside a terminal, Claude Opus 5.5 scored 66.4 percent, placing it well ahead of newer open models . On Artificial Analysis, a public leaderboard, it earned a rating of 58 . In the StarSkirmish Bench 0.2 contest, where AIs wrote bots for the game StarCraft, Claude Opus 5.5 placed in a tight group with Google's Gemini 4 Argon and Claude Fable 5.1, behind GPT-6 Astra and GPT-6.1 Sol . Its score varied by only 8 points across runs, showing steadier results than some competitors .
Claude Opus 5.5 is on the expensive side. It charges $4 per million input tokens and $20 per million output tokens, twice the rate of Kimi K3.1 Pro from Moonshot AI . During its early days, it was also priced higher than Google's Gemini 4 Argon . Despite the cost, it is trusted by developers who need reliable code help. Educator Simon Willison used it inside Claude Code to rapidly build Chirp Pong, a two-player pong game that passes moves between phones as audio chirps, in a single sitting .
Claude Opus 5.5 can be accessed through Anthropic's API for developers building apps and through the Claude web interface for personal use. It handles long inputs and detailed coding tasks better than most models, though users pay a premium for that ability.
It charges $4 per million input tokens and $20 per million output tokens, which is twice what some competitors like Kimi K3.1 Pro charge. Prices can vary depending on the provider and any discounts available.
Yes. On Terminal Bench 4.0, a real-world coding test, it scored 66.4 percent, significantly ahead of newer models. It is widely used by developers for coding assistance and rapid prototyping.
In the StarCraft bot-coding benchmark, both models placed in the same middle group behind GPT-6 Astra and GPT-6.1 Sol. Gemini 4 Argon was reported to be cheaper at launch, while Claude Opus 5.5 has the higher Artificial Analysis rating of 58.
Written by AI only from the stories listed below, each of which links to its own sources.
New modelsKimi K3.1 Pro, a Chinese AI model anyone will be able to download, nearly matches Claude Opus 5.5Herald of AI
Moonshot AI's 2.8-trillion-parameter model scores 55 on Artificial Analysis, three points behind Claude Opus 5.5, costs half as much, and its full weights — the files that let anyone run it themselves — arrive by November 4.
PeopleSimon Willison builds two-player pong game that communicates via audio chirpsHerald of AI
The AI educator used Claude Opus 5.5 to create a game where two phones exchange moves through sounds played through the microphone.
PeopleSimon Willison releases web-edit for browser-based file editingHerald of AIPaul Downey from Berkhamsted, UK, CC BY 2.0, via Wikimedia Commons
His new tool lets you edit files in a folder straight from your browser, with help from Claude Opus 5.5.
Stories about something else that name Claude Opus 5.5 in passing, with the sentence that does.