… The company states that you can give this new model a try right now; sure enough, several members of the XDA team report that they can select the model on the Claude website, which sports the description "Most capable for ambitious work." If you want an idea of how fast the world of AI is moving, j… …
… Just in the past few days alone, the company launched Claude Design, routines in Claude Code, Claude for Word, Managed Agents, released Cowork generally and the list goes on. For many, the most anticipated release in the entire wave was Anthropic's newest flagship model, Opus 4.7 . …
… The vision upgrade is probably the most underrated part of the Opus 4.7 release for anyone who isn't a developer. Prior Claude models processed images at around 1.15 megapixels - Opus 4.7 handles up to 3.75, which is more than three times the resolution. …
… I set both up in the Pi harness, ran the same coding test I'd already used against Claude Opus 4.6 as my baseline, and threw Qwen3-Coder-Next into the mix on my ThinkStation PGX as a local reference point. I used Opus 4.6 here because, at the time of testing, Opus 4.7 hadn’t been released yet. …
… To give you an idea of just how much cheaper it is, here are the costs for Anthropic's newly-released Claude Opus 4.7: Model Base Input Tokens 5m Cache Writes 1h Cache Writes Cache Hits & Refreshes Output Tokens Claude Opus 4.7 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok In other word… …
… To some degree, that makes it feel similar to how fast Opus is on Claude Code, since the cloud model pauses to 'think' before answering. Related I finally found a local LLM I actually want to use for coding Qwen3-Coder-Next is a great model, and it's even better with Claude Code as a harness. …
… And if things go well, we should see Claude agents learn from past mistakes and reflect on how they can refine their modus operandi to better serve their users. Related Phew: turns out that Claude Code Pro members are keeping Opus after all It was all one big mistake. …
… That put it above GPT-5.4 at 57.7 and Claude Opus 4.6 at 57.3 on the benchmark that's become the de-facto measure of agentic coding. Fourteen days later, Moonshot AI released Kimi K2.6, and it scraped past GLM-5.1 with a 58.6. …
… OpenCode handles multiple models more seamlessly This isn't surprising at all, but given that Claude Code is an Anthropic product, you're limited to Claude models. That's it. No GPT, no Gemini, no local LLMs, nothing. You get access to Claude's Haiku, Sonnet, and Opus models. …
… Most of these tasks didn’t require a frontier model like Opus 4.7 or even Sonnet 4.6. That was when I decided to replace Claude Pro on my mobile with Gemma 4. …