Fact-checking with Claude Opus 5.5
Big If True, our free AI fact-checker, now works with Claude’s newest model. It catches just as many errors, faster and at lower cost, and it now updates itself. Get it at verso.ink/big-if-true.
Last week, we released Big If True, our free fact-checking skill for Claude. This week, Anthropic released a new model, Claude Opus 5.5. It took the top spot on the Artificial Analysis Intelligence Index, an independent ranking of AI models built from many benchmarks.
So we tested it. We ran Opus 5.5 through the same tests we used for the launch: published articles with known errors, and drafts where we planted mistakes. At High thinking effort, it caught as many errors as Fable 5.1 at Medium, finished faster, and cost about a third less to run.
Opus 5.5 is faster and cheaper
Minutes per check
Cost per check
Our tests also turned up one weak spot: Opus 5.5 sometimes dropped claims from its report. To fix this, we designed a new version of Big If True that won’t finish until every single claim has a verdict.
The new version is out now, and it’s easier to keep up to date: Big If True is now a Claude plugin. Instead of uploading a file for every release, you add the plugin once, and updates reach you automatically. Open Big If True in Claude, then click Sync and Add.
Using the old skill? Once the plugin is installed, remove the skill under Customize → Skills, so Claude always uses the new version (currently 0.7.0).
When you start a check, we recommend using Opus 5.5 at High thinking effort.
As always: Big If True can still get things wrong, so look at the sources before you change anything. If it catches something, let us know at [email protected].
Happy fact-checking!