Flipping high-value items
At the top of the market the buy limit is tiny and rarely the constraint, the tax finally gets cheaper, and the only two things that decide whether you make gold are how much you paid to get in and how long you are willing to wait to get out.
On 24 October 2026 a twisted bow traded around 1.47 billion gold and changed hands roughly 480 times a day. An elysian spirit shield sat near 400M on about 380 daily trades. A third age longsword was worth around 755M to 796M, and it traded about a dozen times in twenty four hours. This is the thinnest, slowest and most expensive corner of the Grand Exchange, and it plays by rules the rest of the market never sees.
What counts as high value
For flipping purposes the line is not a fixed number, it is a behaviour. An item belongs in this lesson when its daily trade count is small enough that individual buyers and sellers matter. A dragon warhammer at 16M trades nearly 1,500 times a day and behaves like normal gear. A torva platebody at 174M trades 230 times a day and does not. The price tag is a symptom; the thinness is the disease.
Practically, that means anything from roughly 100M upward, plus a handful of cheaper items with unusually few trades. Ancestral robe tops near 86M, zaryte crossbows near 360M, elysian shields near 400M (24 October 2026), torva pieces, third age gear and the twisted bow all sit here.
The tax cap, and where it starts helping
The Grand Exchange sales tax is 2% of the sale price, capped at 5,000,000 gp per item. For every category earlier in this track that cap was irrelevant, because 2% of the sale price never got near it. Here it becomes the single most favourable rule in the game.
The cap starts binding at a sale price of 250,000,000 gp, because 2% of 250M is exactly 5M. Above that, the effective tax rate falls away.
Sell something at 500M and the tax is 5M, an effective rate of 1%. Sell at 1B and it is 0.5%. Sell a twisted bow at 1.47B and the 5M fee is about 0.34% of the sale. The most expensive items in the game are also the cheapest to sell, and that is a genuine structural advantage for anyone with the bank to reach them.
Slippage: the screen price is not a price
This is the part that costs people the most gold, and it deserves more attention than the tax.
On a liquid item, the quoted buy and sell prices are meaningful, because thousands of trades a day keep them honest. On an item that trades ten times a day, the quoted price is a summary of a handful of transactions that may have happened hours apart, at prices that differed by a large amount. A third age longsword with a recent range of 814M to 1.13 billion has a band nearly 40% wide. Any single number you see for that item is close to meaningless as a guide to what you can actually transact at right now.
Slippage is the gap between the price you expected and the price you got. In this category it comes from three places. Your buy offer may only fill by paying more than the last trade, because the seller who was there has gone. Your sell offer may only fill by accepting less, for the same reason in reverse. And if you try to move more than one unit, each additional unit costs you more than the last, because you are working through a very short queue of counterparties.
Assume slippage before you assume profit. If a flip only works when you buy at the bottom of the recent band and sell at the top of it, it is not a flip. It is a bet that two rare events happen in your favour.
Capital lock-up
A twisted bow position is about 1.47 billion gold sitting in one offer on one item. While it sits there it earns nothing, it cannot be redeployed, and it cannot be partially retrieved if you change your mind. Compare that with the coal flip from lesson four, which committed 1.79M and closed within the hour.
The honest way to evaluate a high-value flip is against what that same gold could have earned elsewhere over the same period. If 1.47 billion could have run twenty simultaneous positions in liquid consumables all evening, then a single 3% flip that takes three days to close needs to beat that, not just be positive. Very often it does not.
This is why experienced flippers with large banks usually keep only a portion of their gold at the top of the market, and run the rest through faster categories. It is also why a player who has just crossed 100M should be careful about immediately putting all of it into one prestigious item: the return on your first big flip matters far less than the flexibility you gave away to make it.
Patience is the strategy
Everything above adds up to one behaviour. In this category the flipper who wins is the one who sets a price and waits.
Waiting is what earns the spread. There is no crowd of counterparties competing to fill your offer, so the profit is the compensation for being the person willing to hold the position while nobody else was. If you cut your sell price after a few hours because nothing happened, you have paid for the waiting and then given away the reward. That is the most common mistake here by a wide margin.
So set expectations accordingly. A high-value flip is measured in days, not sessions. Place the buy at a price you would be glad to own the item at, place the sell at a price you would be glad to receive, and then leave both alone. Check once or twice a day, not once an hour. If the fundamentals change, a new best-in-slot announcement for instance, act decisively. Otherwise, do nothing, which is harder than it sounds.
Two worked flips
Above the cap. Twisted bow, limit 8, recent band roughly 1.31B to 1.47B (daily averages, 90 days to 23 October 2026).
Buy at 1.35B, sell at 1.40B. Tax = capped at 5,000,000 gp, not 2% (which would be 28M).
Net = 1,400,000,000 − 1,350,000,000 − 5,000,000 = 45,000,000 gp.
Return on capital: about 3.3%, on one unit, over what may be several days.
Below the cap. Ancestral robe top, limit 8, recent band roughly 80.6M to 115.4M (daily averages, 90 days to 23 October 2026).
Buy at 92M, sell at 97M. Tax = 2% of 97M = 1,940,000 gp.
Net = 97,000,000 − 92,000,000 − 1,940,000 = 3,060,000 gp.
Return on capital: about 3.3%, on 92M rather than 1.35B.
The rates match, which is the point. Fifteen times the capital bought you fifteen times the gold and no improvement in return. What the twisted bow did give you was a far gentler tax bill in percentage terms, and what it took was flexibility: 1.35B in one offer against 92M in one offer is a very different amount of your bank to have frozen while you wait.
Platinum tokens and offers over 2.147 billion
Since 30 September 2026 you can place a Grand Exchange offer above the old 2,147,483,647 coin limit, with the payment taken from your coins and your platinum tokens together. A platinum token is worth a fixed 1,000 coins and you swap coins for tokens at any banker, so a bank holding tokens carries far more value than a coin stack can. Offers on the most expensive items, such as the 3rd Age pickaxe at around 8.6 billion on 24 October 2026, can now be placed in full.
For a flipper the rules around that barely move. The tax is still capped at 5 million per item and buy limits are unchanged, so these remain the cheapest items in the game to sell as a percentage. What does change is the guide price: since the same update it refreshes every three hours and can move up to 20% per update, and rarely traded items now update more often too. Count tokens at 1,000 coins each when you price an offer, and be careful in player-to-player trades, where a handful of tokens can be passed off as a large coin stack. Sources: Jagex's "GE Improvements: Beyond Max Cash" post (30 September 2026) and the OSRS Wiki.
Bonds: the accessible end
One high-value item behaves nothing like the rest. Old school bonds traded around 9.9M on 24 October 2026, carry a four hour buy limit of 100, and are explicitly exempt from the Grand Exchange tax. That combination is unique: real size available, no tax at all, and a price band that ran from about 9.8M to 12.2M on daily averages over the 90 days to 23 October 2026.
Bonds also have a demand source outside the game economy, since players buy them with real money and sell them for gold. That gives the price a floor and a ceiling driven by things no in-game update controls, and it makes the item behave more like a currency than like gear. If you want exposure to large positions without the thinness of the top end, bonds are the obvious place to look first.
Risks
- Announcement risk. A single developer blog can move a 500M item by tens of millions before anything trades. You cannot react faster than the news.
- Slippage on exit. Getting in is usually easier than getting out, because sellers are more motivated than buyers at the top of the market.
- Capital lock-up. Gold in a slow offer is gold not working. Measure the flip against what else that gold could have done.
- Wide bands mislead. A price near the bottom of a 40% range is not evidence of a bargain, it is evidence of a thin market.
- Concentration. One item, one position, one story. There is no diversification in a single high-value flip.
Finding candidates
Use the Margins board (free account required) with Buy price over 50M, and treat the Last buy and Last sell columns as the first thing you read rather than the last. On thin items those two numbers are the difference between a live quote and a memory. Sort by ROI high to low, since limits are uniformly tiny here, and add Trades an hour as a second sort so the items that at least move occasionally sort above the ones that do not.
The Flip Finder is useful mainly as a discipline device at this level. Set your capital to the amount you are genuinely willing to freeze in one position for several days and let it size accordingly. If the number you type makes you uncomfortable, that discomfort is information.
Then set the position up so it does not need you. Track it and it becomes an open position in My Flips (free account required) with its running profit and loss, and attach a price alert at the level you decided on before you bought. At this end of the market that is not a convenience, it is the whole discipline: a position measured in days and hundreds of millions is exactly the kind you will otherwise sit and watch, and watching is what turns a planned exit into an impatient one.
That completes the track. You have now seen eight categories that span four gp arrows to 1.4 billion gold bows, and the same three questions decided every one of them: who is buying, how often does it trade, and what does the limit let me do about it. Take those questions back to the Flip Finder or the Margins board and the live market will answer them for you.
High-value pages to watch
Live prices, volume and buy limits for the items named above: Twisted bow, Elysian spirit shield, Ancestral robe top and Old school bond.
GE Uncut is an unofficial, fan-made tool, not affiliated with or endorsed by Jagex Limited. RuneScape and Old School RuneScape are trademarks of Jagex Limited; in-game content and item images are the property of Jagex Limited, used for reference only. Price data is from the OSRS Wiki real-time prices API. See data attribution. Prices and volumes quoted in examples are snapshots, from August 2026 unless a date is shown, and have since moved; item pages show live figures. These guides are written with AI assistance and corrected when errors are found. Where a figure carries a date, it was taken from GE Uncut's own data or the OSRS Wiki on that date. How it works explains how signals are graded and which parts of the site use AI.