Most Dangerous Mountains to Climb: The World’s Deadliest Peaks

By the measure most sources use, the ratio of deaths to successful summits, Annapurna I in Nepal is the most dangerous eight thousander for which current, consistent data exist. The latest Guinness World Records entry, based on Himalayan Database figures as of February 2025, lists 75 deaths against 559 ascents, a rate of 13.42%. An older tally covering all fourteen eight thousanders, published by NASA Earth Observatory with data to March 2012, also puts Annapurna first at about 32%, followed by K2 at about 26.5% and Nanga Parbat at about 20%. Dhaulagiri I is now almost level with Annapurna in the Nepal data.

That qualifier matters. “Dangerous,” “deadliest” and “hardest” describe three different things, and most rankings blur them. A mountain can be brutally hard yet statistically safer than a low, crowded peak that kills the unprepared. Datasets also cover different peaks and different years, so a percentage means little until its source, denominator and time period are stated. This guide states them throughout.

The sections below explain how danger is measured, profile ten of the most serious peaks, and compare K2 with Everest and with Annapurna. They then cover Pakistan, Gilgit Baltistan, Europe and the United States, and end with why so many deaths happen on the way down. The guide also covers the July 2026 avalanche on Broad Peak.

What Is the Most Dangerous Mountain to Climb?

Annapurna I (8,091 m / 26,545 ft) has the highest ratio of deaths to summits among the Nepal Himalaya eight thousanders in current Himalayan Database and Guinness figures, and it also led the older tally that covered all fourteen. K2 (8,611 m / 28,251 ft) is the peak many climbers and analysts rate as the most serious overall once technical difficulty, weather and rescue limits are combined. Mount Everest has the highest total death count among the eight thousanders, but a much lower ratio.

Neither answer is wrong, because they answer different questions. Annapurna’s reputation comes from avalanche exposure and a relatively small number of successful summits. K2’s comes from sustained technical climbing, violent weather and almost no rescue margin above base camp.

Most confusion about the world’s most dangerous mountain comes from mixing up these terms:

Term	What it actually measures	Typical example
Dangerous	Overall exposure to serious hazards, including avalanches, weather, remoteness and route conditions	K2, Annapurna I
Deadliest	A mountain’s recorded deaths or its fatality statistics, depending on the metric used	Annapurna I, Nanga Parbat
Hardest	Technical, physical and logistical difficulty of completing the climb	K2, Baintha Brakk
Highest fatality rate	Deaths divided by a clearly stated denominator, such as successful ascents, climbers or expedition members	Annapurna I
Most deaths	The highest absolute number of recorded climbing deaths, without adjusting for the number of climbers or ascents	Everest among the eight thousanders

Almost every disagreement about the deadliest peak in the world comes down to which row someone is using.

How Do You Measure the World’s Most Dangerous Mountains?

There is no single objective ranking. Different databases use different time periods, different denominators and different rules about who counts as a climber, and no one dataset covers every major peak. Three kinds of source appear throughout this guide.

The Himalayan Database is a set of expedition records for the Nepal Himalaya, first compiled by the journalist Elizabeth Hawley and the researcher Richard Salisbury. It covers Everest, Lhotse, Makalu, Cho Oyu, Kangchenjunga, Dhaulagiri I, Manaslu and Annapurna I, but not the Pakistan peaks. Guinness World Records draws on it for its Nepal comparison.

The March 2012 tally counts ascents and deaths for all fourteen eight thousanders, published by NASA Earth Observatory, which credits statistics compiled by Eberhard Jurgalski. It is dated, but it is the one widely cited source that treats every eight thousander the same way.

For Pakistan and the Karakoram, no synchronized public database equivalent to the Himalayan Database covers K2, Nanga Parbat, Broad Peak and the Gasherbrums, so figures for them come from varied compilations and differ by counting rules.

Any percentage you see is only as meaningful as the method behind it.

Deaths per successful summit. This is the headline metric behind “one death for every three summits.” It divides all recorded deaths on a mountain by the number of successful ascents. In the March 2012 tally, Annapurna I ranks highest of the fourteen eight thousanders, followed by K2 and Nanga Parbat, then Dhaulagiri I and Kangchenjunga. The metric has real flaws: it ignores climbers who turned back safely, it counts deaths of support staff, and it swings wildly when the number of summits is small. It is a useful comparison tool, but it is not your personal odds of dying. A ratio of 13% does not mean that 13% of climbers die.

Fatality rate per climber. A fairer measure of individual risk divides deaths by everyone who set out, not just those who reached the top. The current Himalayan Database report lists 75 Annapurna deaths among 2,059 expedition members, or 3.64%, the highest per member rate among the Nepal eight thousanders in that report. That figure is far smaller than the summit based ratio, yet still very high compared with most mountain objectives. Small differences between sources are normal: the live database report shows 557 Annapurna summits (13.46%), while Guinness lists 559 (13.42%), because of cutoff dates and counting rules. Even the per member figure has limits, since it includes Sherpas and other high altitude workers, who spend far more time in hazardous terrain than clients do. For comparison, a 2026 study in The Journal of Physiology by Paul Firth and colleagues, using Himalayan Database records, found that the overall mortality rate among Everest mountaineers above base camp fell from 1.4% for 1921 to 2006 to 0.7% for 2007 to 2024. Your own risk depends on your route, season, team, oxygen use and judgment.

Total deaths. Absolute numbers reward popularity. Everest, Mont Blanc, the Matterhorn and Mount Washington record many deaths largely because so many people climb them. Total deaths tell you where the most people have been lost, not where each climber is most at risk.

Technical difficulty describes how demanding the climbing is: steep rock and ice, mixed terrain, route finding and sustained exposure. It overlaps with danger but is not the same thing. Cho Oyu has one of the easiest standard routes of any eight thousander and one of the lowest death ratios, about 1% in the March 2012 tally and about 1.3% in the current Himalayan Database report.

Objective hazards are dangers that skill cannot fully remove, such as serac fall, avalanches, rockfall and sudden storms. A strong climber can be killed by a falling block of ice simply for being in the wrong place. This is why some statistically modest mountains still feel brutal to those who climb them.

Rescue and remoteness determine how much a mistake can be recovered from. Helicopters struggle at extreme altitude, weather can ground them for days, and rescue above 8,000 meters is largely a matter of teammates helping each other. Remoteness turns injuries into fatalities.

Expert tip: When you see a fatality percentage, ask three questions. Whose deaths were counted, what was the denominator, and which years does the data cover? A figure that ignores any of these can mislead by a factor of two or more, as the two Annapurna figures above show.

10 Most Dangerous Mountains to Climb

This list is an editorial judgment of overall seriousness, not a scientifically calculated or strictly fatality based ranking. Fatality data are one input, and where they exist they are given below with their source and period. They do not come from one dataset, because none covers all ten peaks consistently. The order also weighs technical difficulty, objective hazards such as avalanches and serac fall, remoteness, weather and the realistic prospects of rescue. No formula or weighted score sits behind it, and reasonable analysts could order some of these differently.

The top four follow the fatality evidence and hazard profile closely. Positions five to ten are a looser group. Dhaulagiri I, for instance, has a current ratio almost level with Annapurna’s yet sits fourth, because K2 and Nanga Parbat add greater remoteness, weaker rescue prospects and, in K2’s case, far harder climbing. Everest sits eighth despite a low ratio because of its total toll and its specific icefall hazard. Figures will keep shifting as new seasons are recorded.

1. Annapurna I (8,091 m), Nepal

Annapurna I was the first eight thousander ever climbed, by Maurice Herzog and Louis Lachenal on 3 June 1950. Herzog lost his fingers and toes to frostbite on the way down, which set the tone for the mountain’s reputation. The danger is dominated by avalanches and hanging ice; Guinness notes that most deaths occur at a single avalanche prone point around 5,900 m, between Camp II and Camp III.

The famous “one in three” figure is dated. Guinness’s 2018 entry, using data to 31 January 2018, put the rate at 27.6%, and the March 2012 tally gives about 32%. The newer 13.42% (75 deaths, 559 ascents, Himalayan Database data as of February 2025) is lower largely because far more people have summited since, which enlarges the denominator, not because the avalanche terrain changed.

One clarification saves confusion: the Annapurna Circuit trek is a different undertaking from climbing Annapurna I. The deadly October 2014 blizzard that killed more than 40 people struck trekkers near Thorong La, and it was a weather disaster on a trekking route rather than a mountaineering accident on the peak.

2. K2 (8,611 m), Pakistan/China

K2 is the world’s second highest mountain and, in the view of many experienced climbers, the most serious of the eight thousanders. The “Savage Mountain” nickname is usually traced to the 1953 American expedition, during which Art Gilkey died. The standard Abruzzi Spur demands sustained rock and ice climbing through features such as House’s Chimney and the Black Pyramid, all at extreme altitude. Its most infamous hazard is the Bottleneck, a narrow gully at about 8,200 meters beneath unstable hanging ice. In August 2008, eleven climbers died after a serac collapse and related accidents on the upper mountain.

K2 lies outside the Himalayan Database, so its figures come from other compilations, and they disagree. The March 2012 tally gives 81 deaths against 306 ascents, about 26.5%. A compilation through 2025 published by Global Summit Guide gives 92 deaths and 964 summits, or 9.54%, while an estimate for August 2023 cited on Wikipedia’s list of deaths on eight thousanders counts about 96 deaths and 800 summits, near 12%. A fair reading is that K2’s current ratio sits somewhere around 9% to 12%, and that the drop from the 2012 figure mostly reflects a much larger denominator, not a mountain that has become dramatically safer.

K2 also became the last eight thousander climbed in winter, on 16 January 2021, by a ten person Nepali team.

3. Nanga Parbat (8,126 m), Pakistan

Nanga Parbat earned the name “Killer Mountain” early. The first recorded attempt on any eight thousander was here in 1895, when Albert Mummery and two Gurkha companions died in an avalanche. In 1934 a German expedition lost ten members, most of them in a prolonged storm during retreat. More than 30 people had died on the mountain by the time Hermann Buhl made the first ascent, solo, in 1953, by a different route.

The Rupal Face rises roughly 4,600 to 4,700 meters from base to summit, depending on the source, and is often described as the highest mountain face on Earth. In the March 2012 tally, Nanga Parbat had 68 deaths against 335 ascents, about 20%, and news reports still commonly describe it as roughly one in five. No synchronized database covers Pakistan’s peaks, so newer figures vary by source and should be treated as approximate.

In February 2016, Muhammad Ali Sadpara, Simone Moro and Alex Txikon made the first winter ascent, an achievement that carries particular weight in Pakistan.

4. Dhaulagiri I (8,167 m), Nepal

Dhaulagiri is the underrated member of this group. Guinness, using Himalayan Database figures, lists 92 deaths against 707 ascents, or 13.01%, nearly level with Annapurna I. The March 2012 tally put its ratio near 15% (69 deaths, 448 ascents). The mountain soars some 7,000 meters above the Kali Gandaki gorge, and its slopes are steep, prone to avalanches and exposed to sudden weather. It was first climbed in 1960 and has never had the media profile of Everest or K2, which is partly why its numbers surprise people.

5. Kangchenjunga (8,586 m), Nepal/India

The third highest mountain in the world is dangerous through length and remoteness as much as technical difficulty. The approach is long, the summit day is very long, and avalanche and weather risks are ever present. The March 2012 tally recorded 40 deaths against 283 ascents (about 14%), while Guinness’s current figure is 7.83%. Climbers traditionally stop just short of the true summit out of respect for local religious beliefs, a reminder that mountain statistics carry cultural context too. Kangchenjunga is not technically extreme, but it remains one of the most demanding eight thousanders.

6. Manaslu (8,163 m), Nepal

Manaslu is now a popular commercial objective, and its cumulative ratio has fallen sharply as summits have multiplied: the March 2012 tally put it near 10%, while the current Himalayan Database report puts it under 3%. Yet in September 2012, an avalanche at Camp III killed eleven climbers in a single morning. The mountain shows why cumulative ratios can hide bad seasons: risk on a heavily traveled mountain is concentrated in specific hazards and specific years.

7. Makalu (8,485 m), Nepal/China

Makalu is a steep, isolated pyramid with technical rock and ice near the summit, sharp ridges and strong winds. It is far less crowded than Everest, and there is little margin for error in the final sections. It is widely regarded as among the most technically demanding eight thousanders on its standard route, even though its fatality ratio is moderate: about 9% in the March 2012 tally and under 6% in the current Himalayan Database report.

8. Mount Everest (8,849 m), Nepal/China

Everest has the largest absolute toll among the eight thousanders, with well over three hundred recorded deaths in Himalayan Database based counts. Guinness places its death to summit ratio at 2.50% (about 3.9% in the March 2012 tally). That gap between total deaths and ratio shows how heavy traffic dilutes the ratio even as the count grows. Fixed ropes, commercial infrastructure and better logistics are commonly credited with the fall.

Its hazards are specific. The Khumbu Icefall is a maze of seracs and crevasses where 16 Nepali guides died in an avalanche in April 2014. Crowding, delays and inexperienced clients add risk in the death zone. The exposure is also uneven: in the 2026 Journal of Physiology update, most climber deaths from 2007 to 2024 (76.5%) happened on summit day, while most Sherpa deaths (82.2%) happened during route preparation, the repeated trips through the icefall and above.

9. Broad Peak (8,051 m), Pakistan/China

Broad Peak looks gentle next to K2, which it faces across the upper Baltoro Glacier, and that impression is part of the danger. A long summit ridge and a false summit exhaust climbers who are already tired and low on oxygen. In March 2013, two of the four climbers who made the first winter ascent died on the descent. Its long term ratio has been moderate: the March 2012 tally recorded 21 deaths against 404 ascents, about 5%.

On 30 July 2026, an avalanche struck ten climbers heading for the summit on the mountain’s normal route, at about 6,600 meters between Camp 2 and Camp 3, around 9:30 in the morning local time. It swept them several hundred meters down the mountain, and all ten were confirmed dead by 1 August. They included the Nepali British mountaineer Nirmal Purja, several Nepali guides, a Pakistani guide and clients from other countries. The Alpine Club of Pakistan coordinated the response, helicopters and ground teams searched from the following day, and teams on neighboring K2 abandoned summit plans to help. Recovery of the dead took weeks. Early reports described unstable weather and fresh snow in the preceding days, but analysis of the accident is still under way, so this guide does not assign a cause.

The disaster was among the deadliest single accidents on an eight thousander in recent decades, close to the tolls of the 2008 K2 disaster and the 2012 Manaslu avalanche, which killed eleven each. It also shows something the statistics obscure: these climbers died on the way up, on the standard route, well below the death zone, to an objective hazard rather than to exhaustion or altitude.

One disaster does not by itself establish a new long term fatality rate. The ten deaths occurred before anyone reached the summit, so they add to the death count without adding to the summit count. Long term ratios span decades and thousands of climbers, so a single avalanche is a sharp jump in the record, not proof that the mountain’s underlying risk has permanently changed. This guide does not calculate an updated Broad Peak ratio, because the 2012 tally and the 2026 accident come from different sources and no standardized public dataset for Pakistan’s peaks exists to combine them. The better lesson is that avalanche exposure on a peak with a moderate reputation can overwhelm even a strong, experienced team.

10. Gasherbrum I (8,080 m), Pakistan/China

Gasherbrum I, also called Hidden Peak, is steep, remote and exposed to avalanches and serac fall in its upper couloirs. Its ratio in the March 2012 tally was about 9% (29 deaths, 334 ascents), well below the top tier, but the isolation of the Karakoram makes any accident harder to survive. Its neighbor, Gasherbrum II, is far more forgiving, at about 2% in the same tally.

Also worth knowing: lower peaks with extreme reputations

Some of the highest ratios of deaths to summits belong to obscure peaks below 8,000 meters that few people have climbed. Minya Konka (Gongga Shan) in China has a reputation for a very high ratio in some tallies, and Gasherbrum IV, Masherbrum and Baintha Brakk in Pakistan are respected as far more technical and unforgiving than most eight thousanders. Those numbers rest on tiny, poorly documented samples, so a handful of deaths can swing a ratio by tens of percentage points, and this guide does not rank them. Baintha Brakk, nicknamed “the Ogre,” was first climbed in 1977, when Doug Scott famously crawled down with two broken legs. It has been climbed only a handful of times since.

The Deadliest Eight Thousanders

There are 14 eight thousanders, all in the Himalaya and Karakoram. They are the peaks above 8,000 meters generally accepted as independent mountains rather than subsidiary summits, and their summits sit in the death zone, where the body cannot acclimatize and every hour costs energy and judgment.

Peak	Height	Country	Main risk drivers
Everest	8,849 m	Nepal/China	Khumbu Icefall, extreme altitude, crowds, sudden weather
K2	8,611 m	Pakistan/China	Steep technical terrain, Bottleneck serac, storms, exposure
Kangchenjunga	8,586 m	Nepal/India	Remoteness, avalanches, long summit day, severe weather
Lhotse	8,516 m	Nepal/China	Steep Lhotse Face, extreme altitude, avalanche and icefall exposure
Makalu	8,485 m	Nepal/China	Steep rock and ice, exposed summit pyramid, severe weather
Cho Oyu	8,188 m	Nepal/China	High altitude, crevasses, avalanches, severe weather
Dhaulagiri I	8,167 m	Nepal	Avalanches, steep terrain, exposed ridges, severe weather
Manaslu	8,163 m	Nepal	Avalanche exposure, crevasses, steep terrain, severe weather
Nanga Parbat	8,126 m	Pakistan	Rockfall, avalanches, severe weather, extreme relief
Annapurna I	8,091 m	Nepal	Avalanches, seracs, steep terrain, severe weather
Gasherbrum I	8,080 m	Pakistan/China	Steep couloirs, seracs, avalanches, remote terrain
Broad Peak	8,051 m	Pakistan/China	Long summit ridge, false summit, avalanches, extreme altitude
Gasherbrum II	8,034 m	Pakistan/China	High altitude, crevasses, avalanches, severe weather
Shishapangma	8,027 m	China (Tibet)	Avalanches, steep upper slopes, exposed ridges, high altitude

Which Mountain Has the Highest Fatality Rate?

Among the eight thousanders, Annapurna I has the highest ratio of deaths to summits in both the older tally of all fourteen peaks and the current Nepal data. Exact figures depend on the dataset, so the table keeps them apart: the middle column uses one method for every peak, and the right column uses the newest figure available and names its source.

Peak	March 2012 tally (deaths / ascents)	Latest published figure and source
Annapurna I	61 / 191 = 31.9%	13.42% (75 / 559), Guinness, based on Himalayan Database data through February 2025
K2	81 / 306 = 26.5%	About 9.5% (92 / 964), compilation through 2025; other published tallies are higher, so figures vary by source
Nanga Parbat	68 / 335 = 20.3%	No single standardized current figure; commonly reported at roughly 20%, depending on the dataset and calculation method
Dhaulagiri I	69 / 448 = 15.4%	13.01% (92 / 707), Guinness, based on Himalayan Database data
Kangchenjunga	40 / 283 = 14.1%	7.83%, Guinness, based on Himalayan Database data
Everest	223 / 5,656 = 3.9%	2.50%, Guinness, based on Himalayan Database data

Reading down the middle column is the fairer comparison, though it is more than a decade old. NASA’s own text rounds K2 to about 29%, but its published counts (81 deaths, 306 ascents) work out to 26.5%, the figure used here. Reading across a row is only directional, because the Nepal rows come from the Himalayan Database, K2 comes from separate compilations and Nanga Parbat has no comparable update.

Two caveats matter. The Himalayan Database covers the Nepal Himalaya only, so K2 and Nanga Parbat need a different source system, and comparing them directly with Nepal figures mixes incompatible datasets. The ratios also shrink over time, mainly because summit counts grow: more successful ascents expanded the denominator, while the avalanche and serac exposure did not disappear. Better forecasting, equipment and support may also play a part, but the numbers alone cannot show that the terrain became safer.

Ratios on small, rarely climbed peaks can top these numbers, but their samples are too small to rank reliably. For the fourteen eight thousanders, Cho Oyu has consistently been the safest, at about 1% in the March 2012 tally and about 1.3% in the current Himalayan Database report.

Is K2 More Dangerous Than Everest?

For an individual climber, K2 is generally considered more dangerous than Everest. Everest has more total deaths because far more people attempt it, but K2 has a much higher death to summit ratio in every tally reviewed here, and much tougher climbing. The gap is about 26.5% against 3.9% in the March 2012 tally, which uses one method for both peaks, and roughly 9% to 12% against 2.5% in newer sources.

Factor	K2	Everest
Fatality metric	About 26.5% in the March 2012 tally; recent compilations vary, roughly 9% to 12%	About 3.9% in the March 2012 tally; 2.50% in the current Guinness figure
Technical difficulty	Sustained steep rock and ice, with technically demanding sections at extreme altitude	Standard routes are less technically demanding, but still involve fixed ropes, steep terrain and complex glacier travel
Altitude	8,611 m	8,849 m
Weather	Severe storms, strong winds and short climbing windows	Strong winds, extreme cold and rapidly changing conditions, especially near the summit
Major route hazards	Bottleneck serac, rockfall, steep traverses and exposure	Khumbu Icefall, crevasses, serac fall, high altitude and crowded sections
Rescue	Extremely difficult on the upper mountain because of altitude, terrain and weather	Difficult at high altitude, although the mountain has more established logistics and rescue resources
Climbing infrastructure	Fewer climbers, fewer fixed resources and less commercial infrastructure	Extensive fixed ropes, established camps, commercial teams and larger support network

The main danger on Everest sits in specific places: the icefall low down and the crowded, oxygen starved summit day. K2’s danger is woven through the whole upper mountain, with no easy retreat and little help if something goes wrong. That difference in how risk is distributed is more useful than any single percentage.

Is K2 Deadlier Than Annapurna?

By ratio of deaths to summits, no. Annapurna I is higher wherever the two can be compared. In the March 2012 tally, which treats both peaks the same way, Annapurna stands at about 31.9% and K2 at about 26.5%. The newer figures, about 13.4% for Annapurna and roughly 9% to 12% for K2, come from different sources, so the gap is directional rather than exact, but the order does not change.

The two mountains are dangerous in different ways. Annapurna’s risk is concentrated in avalanche and serac terrain lower on the mountain, and its standard routes are less technically extreme than K2’s, although its south face is one of the great technical walls. K2 spreads its danger across the whole climb through steep rock, the Bottleneck, storms and isolation, which is why many climbers rate it as the harder objective even though the statistics favor Annapurna as the deadlier one.

So “is Annapurna more dangerous than K2” depends on your definition. It is deadlier per summit. It is not clearly more demanding step by step on the standard routes.

What Is the Hardest Mountain to Climb?

There is no official answer, but K2 is the one most often named among the eight thousanders: its steep, technical terrain, storms and remoteness make it a benchmark for difficulty. If you set aside the 8,000 meter cutoff, Baintha Brakk, Gasherbrum IV and Cerro Torre are often cited for pure technical difficulty.

Hardest does not mean deadliest. Baintha Brakk has few recorded deaths, largely because so few people attempt it. Annapurna’s standard routes are not the hardest technically, yet the mountain kills at a high rate because of avalanche terrain that skill alone cannot avoid.

Route choice also changes the answer. Cho Oyu has one of the easiest standard routes of any eight thousander, yet much harder lines exist on the same mountain. The Eiger has a comparatively straightforward west flank and the notorious north face. When someone names a “hardest mountain,” ask which route they mean.

Most Dangerous Mountains in Pakistan

Five of the fourteen eight thousanders stand in Pakistan: K2, Nanga Parbat, Broad Peak, Gasherbrum I and Gasherbrum II. They are among the most serious mountains anywhere, though not equally lethal, and K2 or Nanga Parbat are usually named as the most dangerous, depending on the measure used. Because Pakistan’s peaks lack a database equivalent to the Himalayan Database, all newer figures for them should be read as approximate.

Peak	Height	Character
K2	8,611 m	Highly technical, severe storms, exposed terrain, remote location
Nanga Parbat	8,126 m	Huge vertical relief, rockfall, avalanches, unstable weather
Broad Peak	8,051 m	Long summit ridge, false summit, avalanche and exposure risk
Gasherbrum I	8,080 m	Steep couloirs, serac exposure, avalanches, remote terrain
Gasherbrum II	8,034 m	Comparatively less technical, but still exposed to altitude, avalanches and severe weather

Several regional factors add to the risk. Approaches are long, the summer weather windows are short, and rescue depends on how high the accident happens. Pakistan Army helicopters can reach the lower mountain, as in the January 2018 Nanga Parbat rescue of Élisabeth Revol, but they generally cannot operate on the upper slopes, and weather can ground them for days. Climbing in Pakistan means accepting that reliance on your own team grows as the altitude does.

Most Dangerous Peaks in Gilgit Baltistan

All five of Pakistan’s eight thousanders lie in Gilgit Baltistan, where the Karakoram, western Himalaya and Hindu Kush meet. The region holds some of the longest glaciers outside the polar regions and a dense field of serious 7,000 meter peaks. Which peak counts as most dangerous here depends on the criteria: by reputation it is Nanga Parbat, by overall seriousness it is generally K2, and by sheer technical severity, unclimbed or rarely climbed giants take the prize.

Beyond the eight thousanders, several peaks deserve mention:

Masherbrum (7,821 m): Technical terrain, avalanche exposure and relatively few ascents.

Baintha Brakk (7,285 m): Known as the “Ogre,” this granite peak has a history of highly difficult climbing and relatively few successful ascents.

Haramosh (7,409 m): Remote, steep and technically demanding, with dramatic vertical relief.

Rakaposhi (7,788 m): Known for its exceptionally large vertical rise from the surrounding valleys and long, steep mountain faces.

Latok Group: A collection of steep granite peaks in the Karakoram, known for technically demanding routes and difficult climbing conditions.

These mountains show why fatality statistics tell only part of the story. Many of the most difficult peaks in the region kill few climbers, simply because almost nobody dares attempt them.

Most Dangerous Peaks in the Himalayas

The Himalaya is not uniformly dangerous, and geography explains part of the difference. Nepal’s peaks generally sit in a region with heavier monsoon snowfall and warmer temperatures, which makes for more avalanche prone snowpack, while the drier Karakoram is colder, more glaciated and more remote.

Season matters as much as location. Most Himalayan climbing happens in the window before the monsoon, in April and May, or the window after it, in September and October. Snow loading, storm systems and afternoon weather change the odds within days, and a mountain that looks calm on a cumulative table can turn deadly in a single bad week.

There is also a human dimension that raw statistics obscure. Nepali guides, porters and high altitude workers bear a disproportionate share of the risk, because they make repeated carries through hazardous terrain; on Everest, as noted above, most Sherpa deaths from 2007 to 2024 happened during route preparation. The Himalayan Database is the main reason reliable Nepal figures exist at all, and it also explains why K2 and Nanga Parbat, outside its scope, are described with looser numbers.

Most Dangerous Mountains in Europe

European peaks are dangerous for different reasons. They are far lower than the Himalaya, but they are close, popular and prone to fast weather changes and rockfall, so they attract many climbers who are not ready for them.

Matterhorn (4,478 m / 14,692 ft): Zermatt’s memorial to unidentified climbers commemorates more than 500 deaths since the first ascent in 1865, when four of the seven climbers died on the descent. Estimates of current annual deaths range from roughly half a dozen to a dozen or more, depending on who counts and how deaths are defined. Falls, rockfall, crowding and fast changing weather are the main causes.

Eiger (3,967 m / 13,015 ft): Its 1,800 meter north face, the “Mordwand,” was first climbed in 1938, and commonly cited tallies put the number of deaths on it at 60 or more. The normal route is much easier, so the mountain’s danger depends almost entirely on the line chosen.

Mont Blanc (4,808 m / 15,774 ft): Western Europe’s highest peak draws roughly 20,000 climbers a year by commonly cited estimates. No reliable cumulative death count exists. Figures near 100 deaths a year are sometimes quoted for the whole Mont Blanc massif, which includes many peaks and activities such as skiing, while other sources give far lower numbers for the summit routes alone. French rescue records cited in public reporting count 102 deaths in the Goûter Couloir area between 1990 and 2017, roughly four a year; that couloir on the standard route is a notorious rockfall corridor.

Grandes Jorasses (4,208 m): Its north face and Walker Spur are among the Alps’ great serious routes, dangerous through steepness, cold and commitment.

The most dangerous mountain in Europe depends on the question. Mont Blanc is often cited for the largest number of deaths, though no verified count exists, the Eiger’s north face for route seriousness, and the Matterhorn for the mix of popularity and objective hazard.

Most Dangerous Mountains in the United States

No U.S. peak approaches the Himalayan ratios, but several are dangerous for very different reasons.

Mount Washington in New Hampshire stands at only 1,917 meters (6,288 ft), yet more than 150 deaths have been recorded there since 1849 in commonly cited tallies. New Hampshire Magazine’s list counted 161 at its last update, including some deaths only indirectly associated with the mountain, and the total keeps changing. The summit recorded a 231 mph wind gust in April 1934, which stood as the world record for surface wind speed until 1996, and conditions can turn from mild to freezing and stormy within hours. Its road and rail access mean many visitors arrive underprepared, which is why it is often called deadly because it is accessible.

Denali in Alaska (6,190 m / 20,310 ft) is dangerous for opposite reasons. The standard route is not technically extreme, but the cold, storms and isolation are severe, and rescue is difficult. Its danger is measured in severity rather than in the crowds that drive Mount Washington’s toll.

What Is the Most Dangerous Peak in Colorado?

There is no official answer, but Capitol Peak and the Maroon Bells are named most often. Capitol Peak is famous for its exposed Knife Edge ridge, and the Maroon Bells have a reputation for loose, crumbling rock. Longs Peak has one of the higher death totals among the state’s 14,000 foot peaks, largely because so many people attempt it.

What Makes a Mountain Dangerous?

Danger comes from a combination of conditions, and the worst mountains stack several at once.

Extreme altitude. Above 8,000 meters, the oxygen supply cannot sustain the body for long. Fatigue, altitude sickness and impaired judgment build quickly, so small errors become dangerous. Altitude makes almost every other hazard harder to survive.

Avalanches and seracs. Avalanches and collapsing ice kill climbers quickly and often without warning. Annapurna, Manaslu, Dhaulagiri, K2’s Bottleneck and the July 2026 avalanche on Broad Peak, at about 6,600 meters, all show how a single failure of unstable snow or ice can kill several people at once. No amount of experience removes this risk.

Rockfall. Loose rock falling from above hits climbers with little warning, particularly on steep faces and gullies. The Goûter Couloir on Mont Blanc and the fractured terrain of Nanga Parbat are typical examples. Speed and helmets reduce the risk but cannot eliminate it.

Crevasses can be hidden under snow bridges and swallow climbers who step on them. Glacier travel requires roped teams and careful route selection, and crevasses are an ever present hazard on glaciated approaches such as those in the Karakoram.

Extreme weather. Storms, high winds and temperature drops can strand climbers in the open. Karakoram storms are often described as arriving quickly and lasting long, and Mount Washington shows that weather can be deadly even on comparatively low mountains.

Steep, technical terrain and exposure. Steep rock and ice require sustained concentration, and a fall can be fatal because of the drop below. The same technical sections that are manageable at low altitude can become deadly at 8,000 meters, when hands are numb and oxygen is scarce.

Remoteness and slow rescue. On a remote mountain, a broken leg can be fatal because help is days away. Helicopters generally cannot fly at extreme altitude in bad weather, and the upper mountain often cannot be reached at all. A team’s own resources are all that stand between an accident and a tragedy.

Crowds and underestimation. Popularity brings its own hazards: traffic jams at bottlenecks, unprepared clients and a false sense of security. Accident reports on Mont Blanc, the Matterhorn and Mount Washington regularly involve people who underestimated a mountain that looked easy. Accessibility can be as deadly as altitude.

Why Is the Descent So Dangerous?

Many climbers assume the summit is the finish line, but it is roughly the halfway point. Everest is the best documented case. A BMJ study of deaths from 1921 to 2006 found that, of 94 mountaineers who died after climbing above 8,000 meters, 56% died during the descent from the summit. The 2026 update covering 2007 to 2024 found that most climbers who die still do so on summit day (76.5%), although mortality during the descent from the summit had fallen. Comparable data are less complete on other eight thousanders, but the pattern is well known among mountaineers.

Several factors combine. Energy and oxygen supplies are low, judgment is impaired by hypoxia, and afternoon weather often deteriorates. Climbers who ignore a turnaround time in order to reach the top spend their reserves before starting the hardest part of the day. The 2013 Broad Peak winter ascent and the Matterhorn’s first ascent in 1865 are classic examples: the summits were reached, and the deaths came afterward. The July 2026 Broad Peak avalanche is a useful contrast, because those climbers died on the way up, to a hazard that better pacing could not have avoided.

Descending also demands different skills. Falling forward on steep terrain is more likely, rappel mistakes are common when tired, and frostbite and dehydration accumulate through the day. On K2, the 2008 serac collapse at the Bottleneck struck climbers who were coming down.

Expert tip: Set a firm turnaround time before summit day and follow it no matter how close the top appears. Experienced mountaineers treat the turnaround time as a rule, not a suggestion, because the summit is optional and the descent is not.

Want to See These Peaks Without Climbing Them?

Trekking offers a way to be close to some of these mountains without attempting a summit. In Pakistan, the Baltoro Glacier trek to Concordia and K2 Base Camp puts K2, Broad Peak and Gasherbrum IV in view, and the Fairy Meadows trek offers close views of Nanga Parbat. In Nepal, the Annapurna Base Camp trek and the Annapurna Circuit deliver the scenery without technical climbing.

These are not risk free. High altitude trekking carries altitude sickness, weather and glacier hazards of its own, so acclimatization, a qualified guide and honest fitness planning matter.

FAQ