Anthropic · 2026-07-24 · seismic
Claude Opus 5 — Anthropic's new Opus nears Fable 5 at half the price
Claude Opus 5 is Anthropic's new flagship Opus, priced the same as Opus 4.8 ($5/$25 per 1M tokens) but roughly double the performance on Frontier-Bench v0.1 and three times the next-best model on ARC-AGI 3.
Anthropic's new Opus tier comes close to Fable 5 while holding the Opus 4.8 price.
Key specs
| Price (input) | $5 |
|---|---|
| Price (output) | $25 |
| Hn points | 278 |
Quick facts
| Maker | Anthropic |
|---|---|
| Model ID | claude-opus-5 |
| Price (input) | $5 / 1M tokens |
| Price (output) | $25 / 1M tokens |
| vs Fable 5 | Half the price |
| Availability | Claude.ai, API, Claude Code, Claude Cowork |
| What's new | Doubles Opus 4.8 on Frontier-Bench v0.1 |
Pricing
| Input | $5.00 / 1M tokens |
|---|---|
| Output | $25.00 / 1M tokens |
| Fast mode · Opt-in higher throughput | 2× |
What is it?
Claude Opus 5 is Anthropic's newest Opus-class model, launched July 24 and immediately live on Claude.ai, the Claude API, Claude Code, and Claude Cowork. Anthropic frames it as a thoughtful and proactive workhorse that lands close to the frontier intelligence of Claude Fable 5 at half the price.
How does it work?
The claude-opus-5 checkpoint keeps the Opus 4.8 price of $5 per million input tokens and $25 per million output tokens, and Anthropic reports it roughly doubles Opus 4.8 on the Frontier-Bench v0.1 evaluation. On ARC-AGI 3 the model scores three times the next-best model, and on Zapier's AutomationBench its pass rate is about 1.5× the next-best system. A Fast mode is available at 2× price for higher throughput.
Why does it matter?
Opus-tier pricing has stayed flat while capability has effectively doubled, so teams already paying for Opus 4.8 get a large step up with no billing change. For agent builders the OSWorld 2.0 result — surpassing Fable 5 at about a third of the cost — makes Opus 5 the new default for long-horizon coding and computer-use agents, and it takes over as the default model on Claude Max.
Who is it for?
Claude API customers, agent builders, Claude Max and Pro users
Frequently asked questions
- How much does Claude Opus 5 cost?
- Claude Opus 5 is priced at $5 per million input tokens and $25 per million output tokens on the Claude API — identical to Opus 4.8 and roughly half the price of Fable 5. Anthropic keeps the Opus-tier price flat while roughly doubling agentic performance, so the cost per unit of capability drops sharply.
- Where can I use Claude Opus 5?
- Claude Opus 5 is live today on Claude.ai, the Claude API (model id claude-opus-5), Claude Code, and Claude Cowork. Anthropic has also made Opus 5 the default model on Claude Max and the strongest model available on Claude Pro.
- How does Claude Opus 5 compare to Fable 5?
- Anthropic positions Claude Opus 5 as a workhorse that comes close to Fable 5 at half the price. Opus 5 lands within 0.5% of Fable 5 on CursorBench 3.2 and beats Fable 5 on OSWorld 2.0 at about a third of the cost, while Fable 5 remains ahead on the most demanding frontier evaluations.
- Is Claude Opus 5 safer than earlier Opus models?
- Anthropic ships Claude Opus 5 with its usual safeguards and specifically notes the model is not state-of-the-art on 'risky, dual-use capabilities' like cybersecurity. Frontier dual-use capacity stays gated in the Mythos 5 configuration rather than shipping in Opus 5 to the general API.
Try it
API model id: claude-opus-5 — https://www.anthropic.com/news/claude-opus-5