Anthropic launches Claude Opus 4.7, still behind Mythos
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic released Claude Opus 4.7, touting better coding, sharper vision, and a self‑checking ability.
- Anthropic acknowledged Opus 4.7 still lags behind the unreleased Mythos Preview model, which is limited to a handpicked group of tech and cybersecurity firms for safety reasons.
- Anthropic displayed benchmark results showing Opus 4.7 outperforms Opus 4.6, ChatGPT 5.4, and Google Gemini 3.1 Pro on several key tasks.
- Anthropic denied that the changes were intended to redirect compute resources to other projects, despite speculation that Opus 4.6 had been “nerfed” to save costs or prioritize Mythos.
- Anthropic added a new “xhigh” effort level between high and max, giving developers finer control over reasoning versus latency on hard problems.
- Anthropic is testing a new “task budgets” system to let developers manage reasoning resources on longer tasks.
- Anthropic plans to use the Opus 4.7 rollout to trial guardrails against cybersecurity misuse, aiming toward a broader release of Mythos‑class models.
Why it matters: Developers gain a more capable Claude model with higher‑resolution vision and finer reasoning control, while Anthropic signals that its flagship still lags behind the secretive Mythos system, underscoring a performance gap that could shape competitive dynamics with OpenAI and Google.
