Xuanyi Geng Lowers 3×3 World Record Average to 3.71 at Deqing Small & Special 2026

Xuanyi Geng (耿暄一) has pushed elite 3×3 speedcubing into a new bracket, setting a 3.71-second world record average at Deqing Small & Special 2026 in Huzhou, Zhejiang, China on April 26, 2026. The final was also a showcase of just how compressed the very top of the event has become: Bofan Zhang took second at 4.38, and Yiheng Wang finished third at 4.62. (worldcubeassociation.org)

The headline number matters because it’s not a single “perfect” solve—an average requires repeatable execution under pressure. Geng’s set included an Asian Record single of 2.80 seconds, but the record average was built on three solves clustered in the mid‑3s after WCA dropping rules. (worldcubeassociation.org)

The record: what Geng actually solved

In the 3×3 final, Geng recorded times of 3.79, (4.33), 3.61, 3.74, (2.80). Under WCA rules for an average of five, the fastest and slowest results are removed, leaving the mean of the middle three solves. That produces an official 3.71 average world record—fast, but also notably consistent. (worldcubeassociation.org)

It’s also an unusually “clean” record set. The three counting solves (3.79, 3.61, 3.74) are all in a tight band, which is exactly what world-record averages tend to require now that top finalists are routinely capable of low‑4 or even high‑3 pacing.

Why averages are the most meaningful 3×3 benchmark

Singles get the spotlight—one scramble, one execution, one number. But a world record average is closer to what competitors and spectators experience in finals: repeated solves with different scrambles, limited reset time, and the need to avoid both big mistakes and small hesitations.

At this level, a 3×3 average isn’t just “turning speed.” It’s a compound of:

• Start discipline: staying relaxed and accurate despite the race‑like tension of a final.
• Lookahead and flow: keeping the solve moving through F2L without pauses that can add half a second by themselves.
• Last-layer efficiency: clean OLL/PLL execution (or equivalent last‑layer solutions) with minimal regrips.
• Error avoidance: one lockup or wrong pair can turn a record pace into an average that’s merely great.

Geng’s 3.71 suggests a solve floor that’s drifting downward: even when the outliers are removed, the remaining solves are still fast enough to set history.

Context: record progression in 2026

WCA competition highlights list Geng’s previous 3×3 average world record as 3.84 seconds, set at Beijing Winter 2026 in Beijing, China on January 11, 2026. In other words, the average record dropped again within the same season—an unusually rapid progression for a benchmark that already sat at the limit of human execution. (worldcubeassociation.org)

The Deqing final also highlights a broader trend that fans have been watching for the past couple of years: the sport’s top end is deep enough that world-record contenders are no longer isolated. A final where three competitors are under 4.7 is not “record day luck”—it reflects a competitive environment where the pace is continually reinforced by high-level training, strong local scenes, and frequent competition opportunities.

What the 2.80 single signals (even though it’s dropped)

Because averages drop the best and worst solves, a standout single doesn’t directly “count” toward the record. But it still says something: 2.80 is fast enough to be an Asian Record single at this competition, and it shows that Geng’s ceiling is well below the 3.71 headline. (worldcubeassociation.org)

For fans, a useful way to interpret this set is:

• The average tells you the baseline Geng can hold across multiple scrambles in a final.
• The single tells you what his best-case execution looks like when everything lines up.

When both numbers are extreme in the same round, it often indicates that a solver isn’t just having a hot day—they’re operating in a performance band that’s been earned in training.

Why this matters for everyday competitors

Most cubers will never chase a sub‑4 average, but record averages still shape the sport in practical ways:

1. They reset what “fast” looks like. As records move, competitors recalibrate goals (sub‑10, sub‑8, sub‑6, and so on) and see what kind of execution quality is possible.
2. They influence practice priorities. At the top, the difference between a 4.1 and a 3.7 average often comes down to eliminating small pauses, optimizing transition comfort, and making last-layer execution bulletproof.
3. They make finals more exciting. A tighter field means more outcomes decided by tiny differences—and fewer “safe” leads.

Even if you’re solving in the 15–30 second range, there’s a lesson in averages: consistency is a skill. Training that reduces big outliers—DNFs, lockups, forgotten algs—usually improves your average faster than chasing a single personal-best scramble.

What to watch next

Two storylines are worth tracking after Deqing:

• Can the new pace hold at major competitions? Records set at single-day meets are still fully official, but big events add pressure, travel fatigue, and deeper rounds. Watching whether finalists keep pushing under those conditions is part of what makes the season fun.
• How low can the average go without a format change? As averages approach the low‑3s, there’s less room for human variability. At some point, the sport may see longer periods where the record holds—not because solvers stop improving, but because the margin for improvement becomes razor-thin.

For now, Geng’s 3.71 stands as the clearest signal yet that modern 3×3 is no longer “sub‑5 at the top.” It’s a discipline where championship finals can realistically be decided in the mid‑3s.

Key takeaways

• Xuanyi Geng set a 3.71 3×3 world record average at Deqing Small & Special 2026 in Huzhou, Zhejiang, China, on April 26, 2026. (worldcubeassociation.org)
• His final times were 3.79, (4.33), 3.61, 3.74, (2.80), with fastest and slowest dropped for the average. (worldcubeassociation.org)
• The round also included a 2.80 Asian Record single. (worldcubeassociation.org)