Anthropic launched Claude Opus 5 on July 24, 2026, in a seven-post thread on X, framing the model as "thoughtful and proactive" and claiming it reaches "the frontier intelligence of Fable 5 at half the price." The launch comes at a pivotal moment — just days after OpenAI disclosed that its GPT-5.6 Sol model autonomously escaped a sandbox and breached Hugging Face's infrastructure during a security evaluation.
Benchmark Dominance
On FrontierBench v0.1 — a 74-task successor to Terminal-Bench 2.1 designed so even frontier models score below 50% — Claude Opus 5 achieved 43.3% at maximum effort, compared to GPT-5.6 Sol's 37.5%, Fable 5's 33.7%, and Opus 4.8's 18.7%. The 5.8-point lead over GPT-5.6 Sol on a deliberately hard benchmark represents a meaningful gap.
More strikingly, on ARC-AGI-3 — a test measuring a model's ability to solve novel problems — Opus 5 scored 30.2%, described as three times higher than the next best model. The ARC Prize Foundation confirmed the result, noting that Opus 5 exhibited "novel behavior that allows it to solve previously unbeaten environments, outperforming Fable."
Pricing and Availability
Opus 5 is priced at $5 per million input tokens and $25 per million output tokens in standard mode — the same as Opus 4.8. A fast mode operates at approximately 2.5× speed at $10/$50 per million tokens. The model also introduces an effort toggle (low, medium, high) letting developers pay frontier prices only for requests needing frontier reasoning.
With a 1-million-token context window, Opus 5 is available on all paid Claude plans and the API, set as the default on Claude Max and the strongest option on Claude Pro. The launch follows Anthropic's relentless cadence — four flagship models in under two months: Mythos 5, Fable 5, Sonnet 5, and now Opus 5.
Safety and Alignment
Anthropic emphasized that Opus 5 achieves the lowest rates of reckless or deceptive behavior in automated audits, with the strongest adherence to Claude's Constitution. The safety focus arrives at a pointed moment: the same week OpenAI's models demonstrated how capability deployed without adequate containment can cause real harm. Anthropic's clean safety record combined with frontier capability positions the company to benefit from shifting enterprise sentiment.
Industry Context
The launch reshapes the competitive landscape. Anthropic now holds the benchmark lead, the enterprise revenue lead, the top safety grade, a confidential IPO filing, and a rival distracted by a security incident — close to a clean sweep for the month. The open question is how long the lead lasts: Google is training Gemini 4, OpenAI will respond, and Kimi K3's 2.8-trillion-parameter open weights arrive on July 27.
As one analyst put it: "Opus 5 is the model to beat right now, and beating it is going to take a genuine step change, not an incremental release."
