The Quick Answer
Bitcoin trades near $77,200 after gaining more than 20% in six sessions, and both the prediction market and our eight-model panel land in almost exactly the same place for December 31: somewhere in the seventies, with no single $5,000 band better than about a one-in-five shot. Where Bitcoin finishes is the part they agree on. The live argument is whether it gets a look at $100,000 along the way, and on that question the panel runs roughly ten points more bullish than the market. The full nine-band board, all eight models, and the one number they fight over are below.
Free: The Weekly PM Market Brief
The 8-model panel's graded record, the week's biggest market-vs-model gaps, and what's spiking next. One email, Sundays.
Unsubscribe anytime, one click. We never share your email.
New to markets that let you trade an outcome directly? Here is how Bitcoin price contracts actually work, including how a price band settles.
The Week That Reset The Question
For most of 2026 Bitcoin was a disappointment. It opened the year at $87,500, touched $97,900 on January 14, and then spent six months grinding lower, bottoming at $57,740 on July 1. As recently as August 16 it closed at $62,819.
Then came the squeeze. On August 19 and 20, a record $2.7 billion of bearish crypto positions were force-closed inside 24 hours, the largest on record since 2021. More than $1 billion of Bitcoin short positions closed in roughly one hour. Short positions accounted for about 92% of the damage, against roughly $257 million on the long side, and 172,108 traders were liquidated. For scale, the October 10, 2025 flash crash was a far larger event overall, roughly $19 billion of positions in a day, but almost all of that was traders betting on the price rising: about $16.8 billion of those, against $2.46 billion on the short side. Measured short against short, last week was bigger.
A short position profits when the price falls. When it is force-closed, the exchange buys the asset back on the trader's behalf. Enough of those at once and the buying itself drives the price up, which force-closes more shorts. That is the machine that ran last week, and it matters for one specific reason: forced buying is not a customer. It buys once and leaves no bid behind it.
Underneath the mechanics sat two real catalysts. The US Treasury doubled the maximum size of its long-end bond buybacks from $2 billion to $4 billion per session, targeting the 10-, 20- and 30-year parts of the curve, with Treasury Secretary Scott Bessent saying he has a "big toolkit" for rising yields. Bernstein strategist Gautam Chhugani tied the move directly to it: "The strong trigger in bitcoin was driven by Treasury's move to buyback bonds at the longer end of the yield curve. We are not macro experts, but we do know bitcoin historically has had a positive reaction to liquidity expansion." Separately, President Trump met Coinbase and Robinhood executives at the White House and pushed Congress to pass a "fair version" of the CLARITY Act, the market-structure bill that would set out how digital assets are regulated and where the line falls between the SEC's jurisdiction and the CFTC's.
Spot Bitcoin funds took in $606 million on Thursday, their largest daily haul since May, with BlackRock's fund accounting for 83% of it. Bitcoin closed Friday at $78,335, up 7.26% on the day and more than 22% across the week, its strongest week since 2024. It challenged $80,000 overnight and slipped back toward $77,000.
So: is that a new regime, or an expensive week that changes nothing about December?
The Board
Kalshi lists 28 separate contracts on this question, one for each $5,000 step. Each pays $1.00 if Bitcoin's settlement price lands inside that contract's own band, and nothing if it does not. The table below groups those 28 contracts into nine wider ranges so the shape is readable on a phone. A price in cents is the market's implied chance, so a range priced at 20¢ means the market gives it about a 20% shot.
| Where Bitcoin Finishes 2026 | Market Price | The Panel | Difference |
|---|---|---|---|
| Below $40,000 | 4.3% | 2.5% | -1.8 pts (panel lower) |
| $40,000 To $50,000 | 4.7% | 5.0% | +0.3 pts (agreed) |
| $50,000 To $60,000 | 8.6% | 11.0% | +2.4 pts (panel higher) |
| $60,000 To $70,000 | 13.9% | 17.0% | +3.1 pts (panel higher) |
| $70,000 To $80,000 | 19.8% | 18.0% | -1.8 pts (panel lower) |
| $80,000 To $90,000 | 20.1% | 15.0% | -5.1 pts (panel lower) |
| $90,000 To $100,000 | 10.9% | 10.0% | -0.9 pts (panel lower) |
| $100,000 To $120,000 | 10.6% | 12.0% | +1.4 pts (panel higher) |
| Above $120,000 | 8.0% | 7.0% | -1.0 pts (panel lower) |
How to read this table. Each market figure is the SUM of the mid-prices of the individual $5,000 contracts inside that range, not a single quoted price. All 28 underlying contracts have both a bid and an ask, so every one is a real two-sided quote rather than a one-sided listing, and their 28 mid-prices sum to 100.9% rather than exactly 100%. That 0.9% excess is an overround, the small amount by which a full set of mid-prices exceeds certainty, and not the bid-ask spread. Each panel figure is the median of the eight models for that range; because the median is taken range by range, the panel column sums to 97.5% rather than 100%. Neither column is rescaled to hide that. Prices fetched August 22, 2026, 12:51 PM ET.
Every price on this page is a live quote as of August 22, 2026, and will have moved by the time you read it.
Read the shape rather than any single row. The market's distribution is a broad hill centred just above where Bitcoin trades today, with roughly 54% of the total weight sitting between $60,000 and $90,000. The panel drew nearly the same hill and shifted it about $2,400 to the left. The market's implied midpoint sits near $79,600; the panel's is $77,250, taken from the eight seats' own median year-end prices rather than from the grouped column above. Neither one thinks last week settled anything.
Every Seat's Number
We ran eight AI models on this market without showing any of them the prices. Each received the same package of fetched data: spot price, realized volatility, the full 2026 high-low range, and the sourced news record of the last six sessions. Then each read the other seven models' reasoning, anonymously, and was invited to revise. The numbers below are their final answers, and the panel blend is the median of the eight seats rather than their average.
| Model | Median Year-End Price | Finishes $70,000 to $80,000 | Finishes Above $100,000 | Touches $100,000 In 2026 |
|---|---|---|---|---|
| Fable | $75,000 | 18% | 14% | 27% |
| Opus | $76,000 | 18% | 19% | 38% |
| Sonnet | $79,000 | 19% | 23% | 45% |
| GPT | $79,000 | 18% | 20% | 38% |
| Gemini | $78,000 | 22% | 18% | 38% |
| Kimi | $76,500 | 18% | 20% | 40% |
| GLM | $76,000 | 17% | 19% | 42% |
| DeepSeek | $85,000 | 25% | 20% | 25% |
| Panel Median | $77,250 | 18% | 19.5% | 38% |
| Kalshi's Price | near spot | 19.8% | 18.6% | 28.5% |
Every seat on this panel is graded against real market settlements — records to date: GPT 84% on 5,658 graded calls · Gemini 86% on 4,423 graded calls · Kimi 82% on 2,414 graded calls · GLM 81% on 2,437 graded calls · DeepSeek 80% on 2,462 graded calls. Recomputed daily; the full scoreboard is public.
These are model estimates, not predictions of fact and not financial advice.
Every number in this piece gets graded in public once the market settles. You can check this panel's history on the full graded scoreboard.
More live boards from the same panel: how high Bitcoin gets in 2026 (above $100,000 trades 28.5¢) · the 2026 high-and-low range board (below $55,000 trades 36¢) · the live year-end price board, rebuilt daily. Prices fetched August 22, 2026.
The Market
| Venue | Kalshi, a CFTC-regulated event-contracts exchange (18+; availability varies by state, see where Kalshi is legal) |
| Event | KXBTCY-27JAN0100, "Bitcoin price at the end of 2026", 28 separate price-range contracts |
| The Contract | Each contract resolves YES if Bitcoin's settlement price falls inside its stated range at the January 1, 2027 measurement |
| Settles On | The CF Bitcoin Real-Time Index, not any single exchange's quote |
| Closes | January 1, 2027, 12:00am ET |
| Prices As Of | August 22, 2026, 12:51 PM ET |
The companion market, KXBTCMAXY-26DEC31, asks a different question: how high Bitcoin gets at any point during 2026. Those contracts resolve YES the moment the index prints above the strike, whether or not the price holds. That distinction is the whole story below.
Where The Panel And The Market Actually Disagree
On the destination, this is close to a tie. The panel's median year-end price is $77,250 against a spot price of $77,220. The market's own centre of gravity sits in the same place. On the single most-quoted threshold, finishing 2026 above $100,000, the market says 18.6% and the panel says 19.5%. Nine tenths of a point apart is agreement.
The gap opens on the path. Kalshi's ceiling market, the one asking whether Bitcoin trades above $100,000 at any point during 2026, is quoted 27¢ bid against 30¢ ask, a midpoint of 28.5¢. The panel's median is 38%. Six of the eight seats landed at 38% or higher.
The reasoning behind that gap is a piece of arithmetic the models argued about openly. For a price that drifts very little, the chance of ever touching a level is roughly double the chance of finishing beyond it, because a price can spike through $100,000 in October and still close the year at $70,000. The panel's own year-end number above $100,000 is 19.5%, and 38% is close to double it. The market's 28.5¢ is only about one and a half times its own 18.6% year-end figure.
The doubling rule is a rule of thumb, though, not a law, and a lower ratio can be perfectly rational. It falls when traders expect the price to drift upward rather than wander, when they expect volatility to fade between now and December, or when they think a Bitcoin that reaches $100,000 tends to keep going rather than round-trip. Any of those would make touching and finishing closer to the same event than the textbook case assumes. So the two ratios are not a right answer and a wrong one. They are two different pictures of how Bitcoin gets from here to December, and only the settled market will say which was closer.
What the panel has done is put a number on its picture and sign it. The difference is 9.5 points on one contract, and it concerns the shape of Bitcoin's path rather than its direction. Worth noting what the book looks like underneath: 27 bid against 30 ask means three cents separate what a buyer pays from what a seller receives, which is a meaningful share of a 9.5-point difference on its own.
The Three Bands That Matter
$80,000 to $90,000, the market's heaviest range and the panel's largest disagreement. At 20.1% this is the single heaviest range on the board, and the panel marks it at 15.0%, the widest disagreement anywhere in the year-end distribution. The panel's case is that pricing this band as the most likely outcome requires Bitcoin to hold most of a six-day, 22% gain for another four months, when a large share of that gain was forced buying rather than demand.
$70,000 to $80,000, where Bitcoin already is. Market 19.8%, panel 18.0%. The two agree. Worth noticing on its own terms: the most likely single outcome for Bitcoin over the next four months, according to both a real market and eight independent models, is that it goes roughly nowhere. Four months of headlines, and the base case is the price you are looking at.
$60,000 to $70,000, the give-back range. Market 13.9%, panel 17.0%. This is where the panel puts the money it took out of the eighties. Bitcoin's 200-day average sits at $68,983 and its 50-day at $64,863, both of which the price reclaimed only last week. Several seats treated that zone as the natural landing place if the squeeze premium bleeds out.
Where The Panel Changed Its Mind
Between the two rounds, four of eight seats revised. The largest move came from Kimi, which cut its median by $3,500 after re-weighting how much of the rally was mechanical:
"B, G, and H independently converged on a point I underweighted: a $2.7B forced-buying event leaves no residual bid, and the tape already confirms it. Price failed the $80,000 challenge and slipped to $77,000, so I raise my squeeze-share estimate from a third to roughly half." — Kimi
Sonnet came down $3,000, and its reason is the sharpest thing any seat said about the difference between a catalyst and a price:
"My $82k median implicitly assumed the Treasury/ETF/CLARITY catalysts added incremental upside on top of spot rather than partly explaining why spot already sits above the pre-squeeze base." — Sonnet
Movement ran both ways. GLM raised its median $2,000 and its above-$100,000 number from 12% to 19%, not because it was persuaded by the bulls but because it caught its own inconsistency: "my previous 12% was statistically inconsistent with the 1-sigma upper bound ($100k) derived from 44% annualized volatility."
DeepSeek held at $85,000, which leaves it $6,000 clear of the next-highest seat and $7,750 above the panel median:
"Forecasters B, C, and G emphasize squeeze mechanics and partial mean reversion, but I maintain that ETF inflows and CLARITY Act passage are structural catalysts still building." — DeepSeek
[Editor's note: the CLARITY Act had not passed at the time of this verdict. DeepSeek is pricing expected passage as a forward catalyst, not describing a completed event.]
Opus also held its median, and explained why the peer agreement it did share was not movement: "B's and C's mechanics are the same read I already had, so agreeing with them is convergence, not movement." Its one revision was upward on the ceiling number, for the same arithmetic reason GLM cited.
The net effect of the round was small on the centre and real on the edges. The panel's median barely budged, from $77,500 to $77,250, while the spread between the highest and lowest seat narrowed from $11,000 to $10,000.
What Would Change The Panel's Mind
Four things, each one checkable, with the direction it pushes the number:
- A CLARITY Act vote in September. Every seat named this as the one catalyst not yet spent. Passage pushes the distribution up and would most likely be the event that makes the ceiling market pay; failure or an indefinite delay caps the upside and hands the argument to the mean-reversion seats.
- Whether ETF inflows become a streak. Thursday's $606 million was one day, and 83% of it came from a single issuer. Two or three more weeks of net inflows converts that from a data point into a trend and moves the panel's median above spot. A reversal to net outflows does the opposite.
- A close back below the 200-day average at $68,983. Several seats treated reclaiming both moving averages as evidence the July low is behind us. Losing that level again would push weight back toward the $60,000 to $70,000 band and below.
- Any second liquidation event, in either direction. The panel's distribution is wide because realized volatility is running near 44% annualized. Another record-scale forced unwind, this time of long positions, would fatten the left tail quickly.
When one of these fires, this page gets re-scored and the panel's numbers change with it.
Settlement Timeline
| Year-End Board Settles | January 1, 2027, 12:00am ET, on the measurement that decides all 28 price-band contracts |
| Ceiling Board Settles | Any time it crosses. A one-touch contract resolves the moment the index prints above its strike during 2026, so it can pay out months before the year-end board is measured |
| Settlement Source | CF Bitcoin Real-Time Index, for both boards |
| Next Known Catalyst | A CLARITY Act vote, expected by the industry in September 2026 |
| Then | Weekly spot Bitcoin fund flow reports, which the panel named as its trend test |
| Then | The final trading days of December 2026, the last window in which a ceiling contract can still be triggered |
| Re-Scored | Whenever one of the four triggers above fires. Prices on this page were fetched August 22, 2026, 12:51 PM ET |
Dark Horses The Panel Won't Dismiss
The tails are where a board like this gets interesting, because a broad hill in the middle leaves the outer ranges priced in single digits.
Below $50,000 (market 9.0%, panel 7.5%). Bitcoin was at $57,740 seven weeks ago. Getting under $50,000 requires roughly a 35% fall from here, which sounds extreme until you line it up against what 2026 has already done: down 34% from the January open to the July low, then up 34% from that low to today. Fable kept the heaviest weight down here, arguing the July low "remains live if the squeeze premium unwinds and macro turns."
Above $120,000 (market 8.0%, panel 7.0%). This range pays if the year-end measurement lands above $120,000, which would put Bitcoin back within striking distance of its October 2025 all-time high of $126,198. Sonnet holds the panel's highest number for the upper tail and justified it on volatility alone: at 44% annualized over 131 days, "the width, not the story, sets the ladder."
The $150,000 ceiling contract. Kalshi quotes 5¢ bid against 6¢ ask for Bitcoin touching $150,000 at any point in 2026. The panel's median is 3.5%. That makes it the only rung on the ceiling ladder where the models sit below the market rather than above it, the mirror image of the $100,000 argument one row down. All eight seats landed at or below 8%.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $188M traded
- 2027 Pro Football Champion · $69M traded
- 2028 U.S. Presidential Election winner? · $61M traded
- 2028 Republican presidential nominee · $59M traded
- Pro Baseball Champion · $58M traded
- Will the U.S. confirm that aliens exist? · $32M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
- GTA 6 Release Date: Traders Are Betting November Holds
- NFL Spread Predictions: The Panel Doubts Almost Every Favorite
- Colorado Vs Georgia Tech Picks: Two Untested QBs, One Team That Can Run
- Oklahoma State Vs Tulsa Picks: Does The No. 1 Offense Show Up On Day One?
- How High Will Bitcoin Go In September 2026? Kalshi Vs. AI
The Bottom Line
Bitcoin had its best week in two years, and it did not move the year-end answer. The market and eight independent models both put the most likely finish within a few thousand dollars of today's price, and both spread the rest of the weight across a range wide enough to contain almost any story you want to tell. That is what 44% annualized volatility and 131 days will do to a forecast.
The one place worth arguing about is the ceiling. The panel puts a look at $100,000 before New Year 9.5 percentage points higher than the market's midpoint, 38% against 28.5%. Much of that rests on a rule of thumb rather than a law: for a price with little drift, touching a level runs at roughly twice the chance of closing beyond it. The market's ratio is lower, which is what you would see if traders expect Bitcoin to trend rather than wander, or expect this month's volatility to fade before December. Both pictures are defensible today, and the settlement will say which one described 2026 better.
Either way, it gets graded in public.
18+. Kalshi is a CFTC-regulated exchange and availability varies by state. Nothing here is financial advice, and prediction-market contracts can settle at zero.
Prices on this page are Kalshi's book. If you also trade on Polymarket, code OS4 gets new users a $20 bonus on a $10 deposit — affiliate link; terms as stated by Polymarket; 18+, availability varies by state.
