← Blog · Sep 14, 2026 · Updated hourly at the source
State of GitHub bounties: what the data actually shows
BountyOS scans Algora, Opire, IssueHunt, BountyHub and TaskBounty every hour. Here's what that scan is showing right now — not a marketing estimate, the actual output of the scanner, snapshotted for this post.
The headline numbers (this snapshot)
| Metric | Value |
|---|---|
| Verified WORK bounties tracked | 149 |
| Combined verified value | $14,414 |
| "Sniper" bounties (<48h old, low competition) | 64 |
| Issues scanned per day | ~6,900 |
| Filtered out as noise (honeypots, dead repos, swarms) | 27% |
Live numbers, not fixed to this exact snapshot — see the current radar for right now.
Where the money actually concentrates
The platform split is the most useful thing in this data, and it's lopsided: Algora accounts for roughly 84% of verified bounty value we track, BountyHub around 16%, and Opire a small fraction despite being a legitimate, still-active platform. This isn't a ranking of platform "quality" — it reflects which platforms currently have the most companies actively funding bounties. If you're only checking one platform, that's the gap you're leaving on the table.
27% is noise — and that's the real cost of manual search
Roughly a quarter of everything the scanner finds gets filtered out before it ever reaches a WORK recommendation — archived repos, farmed "clean" issues from bot-generated test repos, swarms with 20+ competing pull requests, or bounty tags that turn out to be stale once cross-checked against the platform's own confirmation. If you're searching GitHub manually, that 27% isn't a rounding error — it's the exact chunk of your time that goes toward chasing bounties that were never real or already gone.
"Sniper" bounties are a meaningful share, not a rarity
64 out of 149 tracked WORK bounties — over 40% — qualify as "sniper": posted within the last 48 hours with fewer than 3 competing pull requests. The common assumption is that fresh, low-competition bounties are rare and you need luck to catch one. In this snapshot they're closer to half of everything worth pursuing — the actual bottleneck is checking five platforms often enough to catch them before someone else does, not their scarcity.
Methodology, briefly
BountyOS scans all five platforms hourly, cross-verifies bounty tags against each platform's own confirmation (e.g. a BountyHub tag only counts once the bot's comment is actually found on the issue — not just inferred from a keyword match), and scores every result 0–115 on confirmed payout, competing PRs, repo trust/activity, and freshness. "Verified value" only counts bounties that passed that check — it excludes anything still unconfirmed.
See it live
These numbers change every hour as new bounties appear and old ones get claimed or go stale. BountyOS tracks all of it automatically instead of asking you to re-check five sites.
See the current radarSee also: Algora vs Opire vs IssueHunt vs BountyHub vs TaskBounty · Get paid to fix GitHub issues