Gold's October pullback is real, and it is small. On October 7, 2026, one widely used close put spot gold at $4,111.33, down 1.26 percent on the day and about 1.1 percent on the week. A futures series showed about $4,132 on October 8, down from roughly $4,321 on September 25. That is a few percent, not a crash. The crash, if that is the word, was earlier. Spot gold set an intraday record near $5,589 on January 28. A closing high near $5,405 was logged the next day. From those prints to a price a little above $4,100, the metal is down about a quarter. The October dip is a wrinkle on that larger fall.
The headline says the wrinkle creates a buying opportunity, and that five gold mining stocks are the way to take it. The headline is the wrong shape. A buying opportunity is a gap between price and value that the dip itself did not invent. The value in a gold miner, this year, is cash. The cash showed up because the gold price rose much faster than the cost of digging it. That gap was already there at $4,300. It is still there at $4,130, for the mines whose costs stayed in a band near $1,500 to $2,000 an ounce. A 1 percent week does not create it. A 1 percent week does not destroy it either. What the week does is remind you that the gap can shrink, because costs do not fall when the quote does, and because some companies wrote their 2026 plans at $4,500.
So the five names below are not a list to buy. They are five machines to consider, in the plain sense of the word. Look at the margin. Look at whether the cost is rising while the price is falling. Look at whether the plan still works under $4,200. If you cannot say those things, you are not considering a stock. You are considering a mood. This is not a recommendation to buy or sell any security, any fund, or the metal.
Why the cash showed up
People ask why gold miners are generating more cash at higher gold prices. The answer is not magic, and it is not a promise that the cash lasts. A mine's big costs are already in motion before the morning quote. The workforce is hired. The fleet is bought. The royalty is a formula. The sustaining capital that keeps the pit safe does not get a day off because gold had a red session. When the price of the ounce jumps and those costs do not jump by the same amount, almost every extra dollar of price falls through to cash. That is operating leverage. It is the entire bull case for owning a miner instead of a bar.
Newmont's second quarter of 2026 is the clean example, and it should stay an example, not a slogan. Commentary on the quarter noted that the realized gold price rose about 33 percent from a year earlier, while costs applicable to sales rose about 4 percent. Thirty-three against four is the cash. You do not need a spreadsheet to see the direction. The price did the work. The cost structure mostly stood still. Multiply that gap by millions of ounces and the bank account changes in a way a royalty on a bar never will.
The same arithmetic is why the shares fall harder than the metal on a down day. On October 7, one report had the metal fund GLD down about 1.7 percent and the miner fund GDX down about 3.1 percent. Newmont's stock was off about 2.5 percent, to $113.54. Barrick was off about 3.8 percent, to $39.18. The gear that made the cash on the way up takes it back faster on the way down. Anyone who calls the pullback a sale, and does not mention the gear, is selling you only the pleasant half of the machine.
There is a second reason the cash looks large, and it is the starting point. For years, all-in sustaining costs for the big miners lived in a world where $1,800 gold was a good year. A cost of $1,500 against $1,800 is a business. A cost of $1,500 against $4,500 is a flood. The flood is what the last four quarters reported. It is not what a normal decade looked like. If you treat the flood as the new climate, you will pay a price for the shares that assumes the flood. October's job is to ask whether you are paying for the flood or for the business that would still exist if the flood receded toward $3,500, or toward the $4,100 that is already on the screen.
The quarter the cash was measured in
The freshest full scorecard is the second quarter of 2026, when the average gold price in that quarter was about $4,512. Four large miners posted all-in sustaining costs against that average. The figures below come from their releases, as compiled in a September comparison. They are history. They are not a forecast of the December quarter, and they are not a ranking you should buy.
Agnico Eagle reported an all-in sustaining cost of about $1,459 an ounce. Against a $4,512 average, that is an implied margin near $3,053. It was the widest of the four. Agnico's cost also ticked down from about $1,483 in the first quarter. In a year when other people paid a boom tax, this cost went the other way. That is the fact. It is not a crown. A cost can rise next quarter. A jurisdiction can change. A shaft can slip. The margin is a photograph of one quarter.
Newmont reported about $1,621. The implied margin was about $2,891. That is still enormous next to any year before this cycle. It is also the company whose 2026 plan, according to an October 8 note, assumes gold at $4,500. Spot gold this week is below that assumption. The cash in the plan was drawn at a higher price than the price on the page. The gap between $4,500 and a little above $4,100 is roughly $400 an ounce. On a large book of ounces, $400 is not a rounding error. It is the difference between the slide the company showed and the market the company is actually in. Newmont can still be very profitable at $4,100. The point is that "more cash" and "the cash in the $4,500 deck" are not the same sentence.
Barrick reported about $1,866, and an implied margin near $2,646. The cost had been rising, from about $1,708 in the first quarter and $1,581 in the fourth quarter of 2025. A rising cost in a falling price is the combination that compresses the flood. Barrick's margin at $4,512 was still wide. Barrick's margin at $4,110 is wide too, if the cost holds near $1,866. If the cost keeps climbing toward $2,000 and beyond, the share is no longer a pure bet on gold. It is a bet on whether the cost climb stops. The October pullback does not answer that. The next cost print does.
AngloGold Ashanti reported about $2,039, and an implied margin near $2,473. It was the highest cost of the four in that quarter, and the cost had marched up from about $1,720 in the third quarter of 2025. Highest cost does not mean broken. At $4,100, a $2,039 cost is still a margin above $2,000 an ounce, which would have been the gold price itself in a different cycle. It does mean AngloGold has less room if the pullback deepens, and more torque if gold runs back toward $4,500. Torque is not safety. People confuse them when the chart is green. October is not that chart.
Kinross is the fifth, and it is the one that shows the damage a pullback can arrive with. A late-September account of the company's own revisions said 2026 and 2027 production guidance had been cut by about 7 to 8 percent, to roughly 1.84 to 1.86 million ounces a year, and that all-in sustaining costs had been lifted by about 9 percent, to roughly $1,850 to $1,900. Fewer ounces. Higher cost. That is the opposite of the Newmont illustration, where price rose 33 percent and a cost line rose 4. Kinross still has a margin at $4,100 if those new costs hold, and the same account put net cash near $1.9 billion, with a target of returning about half of free cash flow. A strong balance sheet is not a production beat. It is what lets a company survive a guide that got worse. Consider the cash. Consider the cut. Do not let one erase the other.
What "higher prices" no longer means
The question about cash at higher gold prices hides a tense. Higher than what. Higher than 2024, yes. Higher than the second quarter of 2026, no. The quarter that produced those $2,500 to $3,000 margins used a gold price around $4,512. This week's tape is closer to $4,110 to $4,150. The miners are still in a high-price world. They are not in the world of the quarter the market just finished applauding. Anyone who models the next four quarters at the last quarter's price is modeling a rally that has not happened.
That is why the gold price outlook for 2026 cannot be a single cheer. One path is that $4,100 holds, yields stop making highs, and the cash stays large because $1,500 to $2,000 of cost still clears by a wide amount. Another path is that the October dip was the start of a move toward the summer lows, and the $4,500 decks inside company plans become the bull case instead of the base. A third path is January's, in which gold is back near $5,000 or above and the cash flood returns. All three are possible. None of them is an October pullback. The pullback is a few percent. The outlook is the range around it.
Yields are the hinge, as they have been all month. The 10-year Treasury has been near 5.3 percent, close to the highest level since 2002. Gold pays no coupon. At 5.3 percent, the bond you did not buy is a daily argument against the bar. Miners feel that argument twice. The metal price softens, and the discount rate on future cash rises. A dollar of free cash flow in 2028 is worth less at a 5.3 percent 10-year than it was when the 10-year was near 4 percent a year ago. The gold stock market can fall even if the mines keep printing cash, because the market is capitalizing that cash at a harsher rate. A buying opportunity, if the phrase is going to be used at all, would be a share price that fell more than the cash and the discount rate justify. You do not get that from a sector headline. You get it from one company's ounces, cost, and net cash, set next to one share price.
Five margins, not five tips
The search phrases around this tape are loud. Gold stocks to buy. Best gold mining stocks. Best gold mining stocks to buy now. Gold mining stocks to buy. Those phrases want a winner. The tape does not have a winner. It has a spread of costs against one price. Here is the spread, as a way to consider the five, not as an order to purchase them.
Agnico is the low-cost print in the quarter just described. About $1,459 all-in, a cost that fell while others rose, a margin above $3,000 when gold averaged $4,512. At $4,130, if the cost held, the implied margin would still be near $2,670. That is the kind of number people mean when they say the miners are minting cash. It is also a number that depends on the cost holding. Agnico's gift to a reader of this pullback is not that the stock is on sale. It is that you can see, in public, a cost far enough under the new price that a further $200 decline does not turn the mine into a hobby. Whether the shares are cheap is a different question, and it requires the market value, the ounce profile, and the balance sheet, not a paragraph.
Newmont is the scale. It is the largest of the group, the one whose realized price did the 33 percent climb against a 4 percent cost climb, and the one whose plan has been tied to $4,500 gold. Scale means the cash, when it comes, is large enough to fund dividends, buybacks, and projects at the same time. Scale also means a miss is large. If the whole book is being managed to a price the market has already left, the next update has to say so. Consider Newmont if you want to watch whether a giant can re-cut a plan without scaring the cash return. Do not consider it because it fell 2.5 percent on a Tuesday.
Barrick is the cost that has been walking up. From the mid-$1,500s in late 2025 to about $1,866 in the second quarter, the direction is the story. A walk up in cost during a walk down in price is how a flood becomes a river. The river can still be deep. At $4,110 and $1,866, the arithmetic margin is still more than $2,200 an ounce, before you argue about what all-in sustaining cost leaves out. The thing to consider is the next step. If the walk stops, Barrick is a high-margin producer that the market has already started to punish in the share price. If the walk continues toward AngloGold's $2,000 and beyond, the punishment can be about the mine, not only about the gold quote. Those are different investments. October does not tell you which one you are in.
AngloGold is the high-cost end of this particular four, at about $2,039. High cost inside a $4,100 market is not distress. Distress was the last decade, when $2,039 would have been a gold price, not a cost. The use of AngloGold in this list is as a ruler. It shows how far the price can fall before a major miner's reported cost is even in the conversation. From $4,130 to $2,039 is a long way. From $5,589 to $4,130 felt long to the people who bought the high, and it still left this cost untouched. The gold price correction has been violent for late buyers. It has not yet been violent for the cost structure of the majors. That is the fact that makes "buying opportunity" such an easy phrase, and such a sloppy one. The opportunity, if any, is that the shares fell with the metal while the business stayed intact. You still have to check the share, not assume the business being intact means the share is cheap.
Kinross is the reminder that intact is not guaranteed. Production down 7 to 8 percent. Costs up about 9 percent, into a band near $1,850 to $1,900. Net cash still near $1.9 billion. A higher share of free cash flow aimed back at owners. This is what it looks like when the pullback "comes with damage attached," which is how one September note put it. The damage is not the gold price. The damage is the guide. A list of gold mining stocks that ignores a cut, because the stock also fell, is a list of prices. Consider Kinross as the test of whether you can hold two ideas. The margin at $4,100 can still be large. The path to that margin got narrower. Narrower is not the same as closed.
The bar, the fund, and the share
Gold investment is not one object. A bar is the price, minus storage, with no coupon and no mine risk. A fund of bars is the same bet with a fee. A miner is the price, minus the cost, plus the gear, plus the country, plus the balance sheet. In the week of October 7, the bar fell less than the miner fund. That will happen again. It is not a defect. It is the contract you signed when you chose the gear.
Central banks are still in the bar, not in your miner. Kitco reported that China's central bank bought about 21 tonnes in September, the largest monthly addition in three years. Other notes called the September add another month in a long streak. That bid is for gold in a reserve, not for Newmont. It can put a floor under the metal and do nothing for a cost that is rising in Nevada or West Africa. Do not borrow the central bank's conviction and spend it on a share. The bank is not underwriting Kinross's guide.
The gold market outlook, said without a hero number, is a fight between that reserve bid and a bond yield near 5.3 percent. If the reserve bid wins and yields ease, the metal can retrace some of the quarter it lost, and the miners' gear will do more than the metal. If yields push higher, the metal can visit levels that still look fine against a $1,600 cost and look less fine against a plan drawn at $4,500. The October pullback is what the fight looks like on a quiet week. It is not the result of the fight.
When more cash is a trap
More cash at higher gold prices has a failure mode, and the industry has used it before. The cash arrives. The plans expand. The sustaining capital that was deferred gets spent, which is good, and the growth capital that was a dream gets approved, which is sometimes good and sometimes a way to bury the cash in a project that only works at the peak price. Costs then rise because everyone is building at once. The all-in number walks up. Then the price walks down. The cash that justified the plan is the cash the plan already spent.
You can see the early version of that walk in the cost table itself. Agnico's cost dipped. Barrick's rose. AngloGold's rose for several quarters. Kinross reset the guide higher. The boom tax is not a theory. It is a line item. A reader who buys the pullback because "they are throwing off cash" and does not ask where the cash is going will find out in the capital budget, not in the headline. Returning half of free cash flow, as Kinross has pointed toward, is one answer. Building a project that needs $4,500 gold is another. The answers are in the filings. They are not in the fact that the stock was red on October 7.
There is a milder trap, and it is the by-product and the mix. Some of these companies produce copper, silver, or other metals beside gold. A strong copper print can make a gold cost look better. A weak one can make it look worse. All-in sustaining cost is a useful number and a constructed one. It leaves out some things an owner still pays. Use it to compare a company to itself, quarter to quarter. Use it carefully when you line five companies up and pretend the lowest number is the best stock. The lowest number in one quarter is a clue. It is not a conclusion.
What would make the headline true
The headline becomes true, in a narrow way, if three things are true together. First, the business cash at $4,100 is close to the cash the market is refusing to pay for, because the shares fell more than the ounces and the costs justify. Second, the cost walk has stopped, so you are not buying a shrinking margin with a cheerful story. Third, you do not need gold back at $5,500 for the investment to make sense. You only need the price to stay in a range where $1,500 to $2,000 of cost still leaves a surplus you can understand.
That is a high bar, and it is a fair one. It is also not met by "gold fell this week." Gold fell about a quarter from January, and the shares of miners fell with their usual extra force, and the costs, for several of these names, did not fall at all. The result is a sector that is less expensive than it was at the high, and not automatically cheap. Less expensive than a mania is the normal state of the world after a mania. It is the start of research. It is not the end.
The headline is false if you need the January price to return in order to feel solvent. It is false if your company just cut ounces and raised costs and you bought the dip because the ticker was down. It is false if the 2026 plan inside the company still says $4,500 and you did not notice. It is false if you wanted a bar, which has no AISC, and you bought a miner because the miner fell more. More decline is not more value. More decline is more gear.
A way to consider them without a shopping list
If you already own gold stocks, the October print is a chance to rewrite one line on a card. What gold price is in the company's plan. What the last all-in cost was. What the net cash is. What fraction of the cash is being returned, and what fraction is being spent on a future that needs a higher price. If the card still works at $4,000, the week's dip is noise around a business. If the card only works at $4,800, you own a rebound, and you should say so in plain words.
If you own nothing, you do not have to solve that on a Thursday. The gold stock market will still be there after the next cost print, and after the next bond auction. Waiting is not a missed buying opportunity when the move you are being sold is 1 percent on the week and the move that mattered was 25 percent over nine months. The 25 percent is already in the price of the metal. It may or may not be fully in the price of each share. That is the work. The work is not to find five tickers before lunch.
If you want the five on one page, keep them as questions. Can Agnico's cost stay near $1,459, or was the dip in cost a quarter, not a habit. Can Newmont's scale still throw off cash if the plan has to be redrawn under $4,500. Has Barrick's cost walk stopped. Does AngloGold's higher cost still leave a margin you accept if gold is $3,800, not $4,500. Did Kinross's cut change the cash you thought you were buying, or only the story. Five questions. No basket at the end. A basket would average a falling cost with a rising one and call the average a view.
The buyer from last October, and the buyer from January
One more split belongs on the card, because the phrase "pullback" pretends every owner is looking at the same loss. They are not. A buyer who has held gold for a year is roughly flat to slightly ahead, depending on the mark. One futures series put the price about 2 percent above where it stood a year earlier, even after the long fall from January. A metal fund was reported up a few percent over twelve months. The person who bought the late-January high is down about a quarter. The person who bought in the autumn of 2025 never lived through a loss of that size. Calling both of them a dip buyer is a confusion.
The miners split the same way, only with the gear turned up. Someone who bought a major when gold was still climbing through $3,500 has been paid in cash distributions and in a share price that, even after October, may sit far above that entry. Someone who bought the share in the last week of January bought the flood at flood prices. The October decline in the share is a small extra cut on top of a much larger one. Those two owners should not read the same headline and feel the same urge. The first is deciding whether to add to a business that already worked. The second is deciding whether to average down a price that assumed $5,000 gold as a home rather than as a spike.
Averaging down is not a strategy until the cost still clears. It is a habit. The five names in this piece all still clear, on the last reported costs, at a gold price a little above $4,100. That is why the cash story is not dead. It is also why adding shares is a choice about valuation, not a rescue of a broken mine. A broken mine is one whose cost has met the price. None of these five, on the numbers above, is that mine. A fully priced share of a healthy mine can still be a poor purchase. Health and cheapness are different words. October keeps trying to make them the same word.
Free cash flow is the number that settles the argument, and implied margin is not free cash flow. Take the quarterly average price, subtract all-in sustaining cost, and you have a sketch. You do not have the cash that reached the bank. Tax takes a piece. Working capital takes a piece. A project that was approved in the boom takes a piece. A hedge, if any remains, takes a piece. The sketch can say $2,500 an ounce and the bank can say something smaller, and both can be honest. Before anyone treats a pullback as a sale, the right exhibit is the cash-flow statement, not the subtraction. The subtraction tells you the mine is alive. The statement tells you what the owners actually received. Consider the statement. The headline will still be there when you are done, and it will look less like an instruction.
The correction, named properly
Call the gold price correction what it is. From a late-January high in the mid-$5,000s to a price a little above $4,100, the metal has given back on the order of a quarter, with the exact percent depending on whether you mark the intraday spike or the closing high. October 2026, so far, is a small further step, a two-month low in one account, a week of about 1 percent in another. Miners fell more than the metal on the hard day, which is what miners do. The cash they generated at $4,500 gold was real. The cash they generate at $4,100 is smaller and, for the low-cost names, still large. The cash they will generate if costs keep rising and the price keeps falling is the open question. An open question is not a sale.
Why are gold miners generating more cash at higher gold prices. Because the cost did not rise with the price. Newmont's 33 percent against 4 percent is the picture. The picture has a date on it. The date is a quarter whose average price is already above this week's tape. Treat the cash as a fact about that quarter. Treat the October pullback as a small change in the price, not as a coupon someone handed you. Consider the five mines as five costs. Buy, if you ever buy, only when the cost, the plan, and the share price can sit in the same sentence without the word opportunity doing the work.
A note on sources and limits
Price marks for early October 2026 are taken from public market notes, not from a single official fix. GoldPrice.com reported a Wednesday, October 7 close of $4,111.33, down 1.26 percent on the day and about 1.1 percent on the week, and a late-January closing high of $5,405 on January 29, with that close about 23.9 percent under the high. A separate October 8 note put the January 28 intraday record near $5,589.38 and the decline from that print near 26 percent, with an October 7 spot close near $4,109.90, the lowest since August 5. A futures series showed about $4,131.9 on October 8 and about $4,321 on September 25. Intraday quotes on October 8 also printed in the mid-$4,100s. The metal moves during the day. None of these is a forecast.
The same October 8 note reported GLD down about 1.67 percent, GDX down about 3.13 percent to $85.46, Newmont down about 2.45 percent to $113.54, and Barrick down about 3.78 percent to $39.18, and it tied Newmont's 2026 guidance to a $4,500 gold assumption. Those share prices are that session's, not a live quote. The 33 percent rise in Newmont's realized price against about a 4 percent rise in costs applicable to sales is that note's description of the second quarter. It should be checked against Newmont's filing before anyone uses it.
Second-quarter 2026 all-in sustaining costs used here are the figures reported by the companies and lined up in a September 2026 comparison: Agnico Eagle about $1,459, Newmont about $1,621, Barrick about $1,866, AngloGold Ashanti about $2,039, set against a quarterly average gold price near $4,512. Implied margins are that average minus those costs. They are not cash flow, not earnings, and not a prediction. Earlier-quarter costs in the same comparison are included only to show direction. Kinross's production band near 1.84 to 1.86 million ounces, the roughly 9 percent AISC increase toward $1,850 to $1,900, net cash near $1.9 billion, and the free-cash-flow return target are from a late-September account of the company's revised guidance. They are the company's problem to confirm in its own release, and they can be updated.
The 10-year Treasury near 5.3 percent is the early-October 2026 yield discussed in that week's bond coverage. Kitco reported a roughly 21-tonne September purchase by China's central bank, described as the largest monthly add in three years. That purchase is gold for reserves. It is not a bid for any miner named here.
This is not investment advice and not a solicitation to buy or sell any security or commodity. Costs rise. Grades disappoint. Plans get redrawn. Gold can fall further or retrace the January loss. Readers should read the quarterly filings and speak with a licensed adviser before any decision.