Claude Opus 5 is now in Claude Code with Fast Mode (2.5x speed) at Opus 4.8 pricing. Run
claude code --model opus-5to start saving 50% on tokens immediately.
Key Takeaways
- Claude Opus 5 is now in Claude Code with Fast Mode (2.5x speed) at Opus 4.8 pricing.
- Run
claude code --model opus-5to start saving 50% on tokens immediately.
What Changed — Claude Opus 5 Drops at Half the Price of Fable 5
On July 24, 2026, Anthropic released Claude Opus 5, their newest Opus model, and it's available in Claude Code right now. The headline: Opus 5 approaches the frontier intelligence of Fable 5 but at roughly half the cost per task. It's priced identically to Opus 4.8, meaning you get a massive upgrade for the same token spend.
Key stats from Anthropic's announcement:
- New state-of-the-art on several coding and knowledge work evaluations
- Fast Mode runs at 2.5x the default speed
- Lowest rates of reckless or deceptive behavior in automated audits
- Available on all paid Claude plans and the API
This isn't a minor iteration. Opus 5 is the default model on Claude Max and the strongest option on Claude Pro. For Claude Code users, this is the model you should switch to today.
What It Means For You — Concrete Impact on Claude Code Daily Use
If you're using Claude Code with Opus 4.8 or any Sonnet model, here's what Opus 5 changes:
1. Token savings without sacrificing quality. Opus 5 delivers better results per token than its predecessor. For long-running coding sessions — refactoring large codebases, multi-file changes, complex debugging — your cost per task drops significantly.
2. Fast Mode for iteration speed. When you're in a tight feedback loop (e.g., "fix this bug, run tests, fix again"), Fast Mode gives you 2.5x throughput. Use it for rapid prototyping and save full reasoning for the trickiest problems.
3. Better agentic behavior. Claude Code already scores Terminal-Bench 2.1 at 78.9% with Opus 4.8. Opus 5's improved alignment means fewer hallucinated commands, less tool misuse, and more reliable multi-step workflows.
4. Available in GitHub Copilot too. If you switch between Claude Code and Copilot, Opus 5 is available in both. Your prompts and patterns transfer.
Try It Now — Commands and Config to Switch Today
Option 1: One-off use
claude code --model opus-5
Option 2: Enable Fast Mode for rapid iteration
claude code --model opus-5 --fast
Option 3: Set as default in CLAUDE.md
Add this to your project's CLAUDE.md:
# Model Configuration
preferred_model: opus-5
fast_mode: true # enables 2.5x speed by default
Option 4: Per-session toggle
Once inside a Claude Code session, you can switch models with:
/settings model opus-5
Then toggle Fast Mode with:
/settings fast on
When to Use Fast Mode vs Full Mode
| Use Case | Mode | Why |
|---|---|---|
| Debugging a single test failure | Fast | Speed matters, low complexity |
| Refactoring a 500-line function | Fast | Iterative, you'll review anyway |
| Designing architecture for a new module | Full | Needs deep reasoning |
| Code review of a PR | Fast | Pattern matching, not creation |
| Security audit of a codebase | Full | Needs careful analysis |
Pro tip: Start every task in Fast Mode. If the output isn't good enough, switch to Full Mode and re-run. You'll save tokens on 80% of tasks.
What About the New Security Plugin?
Alongside Opus 5, Anthropic released a Claude Security Plugin for Claude Code (beta). It's a multi-agent vulnerability scanner that runs in your terminal. With Opus 5's improved reasoning, the scanner is more accurate and faster. Install it with:
claude code plugin install security
This pairs well with Opus 5 for CI/CD pipelines — run the security scan on every commit.
Bottom Line
Claude Opus 5 is the model upgrade Claude Code users have been waiting for. Same price as Opus 4.8, near-Fable 5 intelligence, and a 2.5x Fast Mode. Switch today and watch your token usage drop.
Source: news.google.com
Originally published on gentic.news
Top comments (0)