Claude Opus 4.8
Opus 4.8
elsewhere's corpus does not cover Claude Opus 4.8 as a subject in its own right. The only direct mentions of it appear in benchmark comparisons from a July 2026 Monolith砺思资本 piece on Moonshot AI's Kimi K3, where Opus 4.8 scored 56 points on an unnamed index (trailing K3 at 57, GPT-5.6 Sol at 59, and Claude Fable 5 at 60) and posted a 14.40x acceleration ratio on the "Mega suite" (behind K3's 18.09x and Fable 5's 18.71x) . The retrieved material about the Claude Opus line itself is limited to the 2024 Claude 3 Opus launch, described by 源码资本 and 蓝驰创投 as Anthropic's then-flagship model with "near-human" comprehension, 200K context length, and pricing at $15/$75 per million tokens input/output . No founding date, architecture details, or positioning narrative for the 4.8 version appears in elsewhere's bundle.
AI-generated — may contain errors, please verify.
Coverage
Moonshot AI's Return to Glory
Good fortune lies ahead.
Moonshot AI K3 Kicks Off a Lightning War on the Frontend
The big one is actually here.
72 Hours After Launch, Moonshot AI K3 Posts Its First Results
Open-Source Players in the First Tier
Using Doubao for work? You'll have to sit at the kids' table.
I held it together.
Benchmark Update for AI Benchmarking: Seed 2.1 Pro Urgently Needs to Escape the Gravity of Mediocrity
Claude's kill threshold has arrived.
ZangAI Benchmark Released: GLM 5.2 Takes First Place, Surpassing Opus 4.8
Not an ad, didn't actually get paid for this.
The Fable 5 Ban: The Script Anthropic Wrote Came Back to Act Against Itself
The most powerful AI model in the United States went offline on Friday. Has the era of universal, undifferentiated access to intelligence ended with Opus 4.8? What impact will this have on future social structures?
