Ask "is CNN biased" or "is Fox News biased" and you'll get strong opinions before you finish the sentence. Ask "is Reuters biased" and you mostly get a shrug — which is odd, because Reuters copy is quietly inside more of your news diet than either of them. Wire stories get republished, rewritten, and aggregated by hundreds of outlets. If Reuters leaned one way, the lean would propagate everywhere.
We've already run this question on CNN and Fox News using story-level data. Reuters deserves the same treatment, because the wire-service answer turns out to be more interesting than "it's neutral."
Why wire services are structurally different
Reuters doesn't primarily sell news to you. It sells news to other newsrooms, which means its product is copy that a left-leaning paper and a right-leaning paper can both run without embarrassment. That business model does real work: it rewards speed, attribution, and flat declarative sentences, and it punishes adjectives. Reuters has also operated under its Trust Principles — a formal independence commitment dating to the 1940s — and its style guidance is famously sparing with loaded labels.
None of that makes bias impossible. It changes where you'd have to look for it. An outlet built to avoid framing can still make choices about which stories exist — and selection is a form of bias that never shows up in a headline. We've written before about bias by omission, and it's the right lens for a wire service.
What the bias charts say
The major rating systems place Reuters at or near dead center, usually with high marks for reliability. We won't relitigate the charts here — the short version is that outlet-level ratings are a starting point, not an answer, for reasons we covered in why media bias ratings aren't enough. A "center" badge tells you about an outlet's average. It tells you nothing about the story in front of you.
So let's look at stories.
Where Reuters shows up in our data right now
As of July 17, there are 18 stories on the Signal/noise feed. Reuters appears in 4 of them — classified center every time:
Netflix's Q2 earnings. Six sources: Engadget and Business Insider on the center-left, Newsweek and Reuters in the center, WSJ and CNBC on the center-right. Notably, our framing analysis found nothing to flag on either side. Earnings numbers are earnings numbers.
Trump Media selling early API access to Truth Social posts. Nine sources on this story, from The New Republic and CNN on the left to the NY Post on the right — and Reuters is the only center-rated source in the set.
Canadian wildfire smoke degrading air quality across 20 US states. Sixteen sources, one of the widest spreads on the feed. Reuters sits in the center column alongside AP, Newsweek, and Sports Illustrated.
The White House teleprompter operator betting scandal. Fourteen sources, with Reuters and The Hill rated center. Our framing note on this story flags the right-leaning press for mostly skipping the detail that prediction market Kalshi caught its own user. It flags nothing in the wire copy.
There's a pattern in those four: money, institutions, verifiable events. And in each case, when our system found framing divergence, it found it in the outlets around Reuters, not in Reuters itself.
The more interesting question: where Reuters isn't
Here's the number that should get more attention: 6 of the 18 stories on our feed right now have no center-rated source at all.
When Zelensky fired the defense minister credited with building Ukraine's drone program, six outlets covered it — WaPo, NYT, and The Guardian from the left, AP and France 24 center-left, the NY Post from the right. No center column. The latest wave of US strikes on Iran: NYT, BBC, France 24, Al Jazeera, Politico, Fox News. Again, nothing rated center. The DHS claim that 278,000 noncitizens registered to vote was covered only by Fox News and the NY Post — a textbook left-side blind spot. The deadly flooding in central Texas ran on NYT and BBC and nowhere on the right.
Why does the center column go empty exactly when stories get contested? Some of it is mechanical — wires prioritize what they can verify, and contested political stories are verification nightmares in their first 48 hours. Some of it is resourcing. But the effect on you as a reader is the same regardless of the cause: on the most politically charged stories, there is often no "neutral wire version" to fall back on. You're choosing between framed accounts, which is exactly when source counts across the spectrum matter more than any single article.
That's the honest caveat on Reuters' center rating. It's earned — but it's earned partly by sitting out the fights.
So is Reuters biased?
On framing, Reuters has the cleanest profile in our current data: where it appears, the divergence our system detects belongs to the outlets around it. When we put Reuters and Al Jazeera side by side on the same Iran strike, the differences were real but lived in emphasis and sourcing, not in adjectives.
On selection, the critiques from both directions deserve a fair hearing. From the right: wire neutrality is a house style, not a worldview, and choices about which stories merit coverage still get made by people. From the left: scrupulous both-sides attribution can flatten stories where the evidence isn't balanced. Both critiques are about judgment calls that "center" badges can't capture — and our snapshot shows Reuters concentrated in business and institutional stories while contested political stories went uncovered by any center-rated source.
If your standard for "unbiased" is never makes framing choices, no outlet clears it — we've made that case in what's the most unbiased news source. If your standard is framing choices rarely detectable at the story level, Reuters clears it about as well as anyone in our data.
How to check for yourself
Don't take the badge on faith. Pick a story where Reuters appears next to rated-left and rated-right outlets — the Trump Media API story is a live example — and read the wire version last. Watch three things: the verbs, who gets quoted first, and what each version leaves out. The wire copy will usually feel flat by comparison. That flatness is the product. Just remember to notice the days the wire version doesn't exist at all — because that absence is data too.