This is ForecastNest’s 4 hours forecast for SOXL, created 2026-08-24 12:28 UTC. The window closed 2026-08-24 16:28 UTC, when SOXL closed at $111.82, down 7.3% from $120.59 when the forecast was made. It traded between $106.00 and $115.05 along the way (-12.1% to -4.6%).
Prediction window: Existing Mode 1 horizon rules apply; the stored expiry is authoritative.
Prediction time: 2026-08-24 12:28 UTC (8/24/2026, 8:28:28 AM EDT).
Valid until: 2026-08-24 16:28 UTC (8/24/2026, 12:28:28 PM EDT). Target session: 2026-08-24 (New York); During session. Lead time to expiry: 4 elapsed hours. Timing: Premarket. Based on stored prediction time, not verified model execution time.
Want to forecast SOXL -- or any other stock or ETF -- with your own model and see how it stacks up against these? Create your own AI model benchmark on ForecastNest.
5 of the 5 models called a direction, and 4 of those calls were right.
Below are the details.
| Model | Range | Move | Confidence | Price it found | Score | Result |
|---|---|---|---|---|---|---|
| ChatGPT — GPT-5 | $116.20–$125.80 | -3.6% to +4.3% | 63% | — | 0.0 | wrong |
| Gemini — Pro | $115.20–$123.80 | -4.5% to +2.7% | 60% | — | 77.2 | right |
| Grok | $114.20–$122.80 | -5.3% to +1.8% | 38% | — | 77.5 | right |
| DeepSeek | $113.50–$122.50 | -5.9% to +1.6% | 58% | — | 81.0 | right |
| Claude — Opus | $115.20–$124.30 | -4.5% to +3.1% | 56% | — | 76.6 | right |
How these models have done on SOXL over 4 hours
This forecast is settled and is included in the statistics below. The remaining records are forecasts created before 2026-08-24 12:28 UTC. Only settled directional calls count toward accuracy.
| Model | Asked | Still open | No call | Called up | Called down | Right |
|---|---|---|---|---|---|---|
| Gemini — Pro | 1 | — | 0 | — | 1 · 1 right | 100% (1/1) |
| Grok | 1 | — | 0 | — | 1 · 1 right | 100% (1/1) |
| DeepSeek | 1 | — | 0 | — | 1 · 1 right | 100% (1/1) |
| Claude — Opus | 1 | — | 0 | — | 1 · 1 right | 100% (1/1) |
| ChatGPT — GPT-5 | 1 | — | 0 | 1 · 0 right | — | 0% (0/1) |
Every published SOXL forecast these models made over 4 hours before this question was asked, on 2026-08-24 12:28 UTC, plus this settled forecast. Forecasts created after this question are excluded. Results reflect the settlements available when this article is generated. A day-ahead call and a week-ahead one are counted separately — they are different questions. Only settled forecasts are graded, so open + no call + up + down adds up to asked. “No call” is a range too centred on the price to be pointing anywhere. Accuracy counts only forecasts that have settled, and the count behind it is shown beside it — a single right call is not a 100% record.
What each model said
ChatGPT — GPT-5
Result: wrong
View full reasoning
1. Market Regime Assessment: The dominant short-to-medium-term regime is a sharp corrective/downtrend rather than a clean range or breakout. SOXL is down 6.5% over 24h, 7.2% over three days, 15.2% over seven days and 37.3% over 30 days. The 4h and daily EMA structures are decisively bearish, OBV is falling, and ADX/DI readings favor sellers. However, the weekly structure remains structurally bullish because the 20-week EMA remains well above the 50-week EMA and the ETF is still up 339.3% over one year. I therefore interpret the current state as a high-volatility correction within a still-positive longer-term trend. 2. Technical Synthesis & Timeframe Logic: For a four-hour forecast, I give the greatest weight to the 1h and 4h charts, with the daily chart acting as a directional constraint and the weekly chart as background context. The 1h picture has the strongest immediate rebound evidence: RSI is 40, stochastic %K is above %D, MACD remains negative but its histogram has risen for three consecutive readings, and price is sitting almost exactly on the 9-hour EMA. The 4h chart is more conflicted: RSI is rising from a depressed level and MACD histogram is improving, but stochastic remains oversold, the EMA stack is strongly bearish, and -DI remains above +DI. This combination favors a tactical rebound attempt rather than a confirmed trend reversal. The daily chart argues against a strong bullish call because its RSI is falling, stochastic is bearish, MACD histogram is deteriorating, price is substantially below both the 20-day and 50-day EMAs, and OBV continues to fall. The weekly chart prevents me from turning the immediate forecast aggressively bearish because its EMA structure remains bullish and stochastic is marginally bullish near oversold territory. 3. Sentiment Context & Integration: The supplied ETF sentiment score of 57/100 is neutral. That is consistent with the technical conflict: the short-term tape is oversold and showing early momentum improvement, while the intermediate trend remains clearly damaged. The lack of a strongly bearish sentiment reading makes a reflexive bounce plausible, but it does not provide enough evidence for a durable reversal. 4. Price Range Justification: I expect the next four hours to trade primarily between the $116-$126 area, with a mild upward skew. $116.71 is the supplied near-term support and is particularly important because it is only 3.2% below the current price. The 1h ATR is 2.9%, while the 4h ATR is an unusually large 8.0%; therefore a four-hour range materially wider than a normal 1h fluctuation is justified. I place the lower bound at $116.20, slightly below support to allow for a liquidity sweep, while $125.80 is below the major 4h resistance at $156.92 but close enough to encompass a rebound toward the 9/26 EMA area. The range is deliberately skewed upward because the immediate momentum evidence is improving after a substantial decline, but the upper bound remains constrained by the strong bearish EMA structure. A sustained break below $116.71 would materially weaken the rebound thesis; a move through approximately $125-$126 would strengthen it. 5. Counter-Argument & Key Risks: The strongest counter-argument is that SOXL is in a genuine momentum-driven liquidation phase rather than merely experiencing a temporary oversold condition. The daily and 4h EMA spreads are large, OBV is falling, and the 4h -DI/+DI relationship remains bearish. Because SOXL is a 3x leveraged semiconductor ETF, an adverse move in semiconductor stocks can translate into an unusually large intraday decline. If $116.71 fails decisively with expanding volume, the $110-$115 region could become the next short-term downside zone. Conversely, the key bullish risk to the bearish medium-term thesis is a sharp semiconductor rebound: the rising 1h MACD histogram, rising 4h RSI and oversold 4h stochastic could produce a fast short-covering move toward $125-$130. 6. Confidence Attribution & Catalysts: Confidence is 63%, above neutral but well below high conviction. The four-hour forecast has a reasonably coherent tactical rebound case, but it conflicts with the dominant daily trend and unusually high 4h volatility. The most important catalyst is the regular-session opening at 13:30 UTC, because the current price is pre-market and extended-hours liquidity is thin; the opening order flow can therefore invalidate the apparent pre-market equilibrium quickly. Beyond the four-hour window, semiconductor-sector news and the broader Nasdaq risk environment are especially important, while the longer-term weekly trend remains a meaningful stabilizing factor. Overall, I favor a modest rebound/consolidation over the next four hours, but not a confirmed trend reversal.
Gemini — Pro
Result: right
View full reasoning
(1) Market Regime Assessment: SOXL is currently entrenched in a short-to-medium-term bearish retracement regime. Across the 1h, 4h, and Daily timeframes, the EMA stacks are decisively bearish (20 < 50), though the long-term Weekly structure retains a bullish posture. The immediate regime is characterized by high volatility and downward drift, punctuated by weak mean-reversion attempts. (2) Technical Synthesis & Timeframe Logic: There is a clear conflict between trend and momentum. The overarching trend (Daily MACD falling, 1h ADX at 32 indicating a strong bearish trend) is downward. However, intraday oscillators show exhaustion: 4h Stochastic is deeply oversold (9/10), and both 1h and 4h MACD histograms are rising. For a 4-hour forecast covering the market open, this conflict typically resolves as a choppy, failed morning rally followed by a resumption of the dominant trend, or a swift rejection at short-term moving averages (like the 1h 26EMA at $124.14). I weight the bearish trend structure higher than the oversold oscillators due to low volume on recent intraday bounces. (3) Sentiment Context & Integration: The ETF Sentiment Score of 57 (Neutral) reflects this exact tug-of-war-bearish structural mechanics battling oversold momentum indicators, creating a highly uncertain short-term environment. (4) Price Range Justification: The 4h ATR is heavily elevated at 8.0% (~$9.64). Given the pre-market timing, widening Bollinger Bands on the 1h and 4h charts suggest an explosive open. My range of $115.20-$123.80 incorporates this volatility, skewing downward to test the established 1h/4h support at $116.71, while capping upside near the 1h 26-period EMA ($124.14), which should act as dynamic resistance to any oversold bounce. (5) Counter-Argument & Key Risks: The primary risk is a 'gap and go' short squeeze driven by semiconductor-specific news or broader Nasdaq strength at the open. If buyers overwhelm the 1h 20EMA ($122.55) early in the session, the oversold 4h Stochastic could trigger a cascade of short-covering toward the $126.79 resistance, invalidating the bearish bias. (6) Confidence Attribution & Catalysts: Confidence is limited to 60% due to the 3X leveraged nature of SOXL combined with the conflicting timeframe signals (oversold short-term vs. bearish medium-term). The opening 30 minutes of regular trading will be the primary catalyst dictating whether the momentum divergence or the structural trend wins the session.
Grok
Result: right
View full reasoning
(1) Market Regime Assessment: Short-term regime is a high-volatility downtrend/consolidation within a broader leveraged semiconductor pullback. 1h ADX at 32 with -DI dominant confirms strong bearish momentum; 4h and daily show weaker ADX (~19) but consistent bearish EMA stacks (20 < 50, spreads 4.7-10.5%) and price below short EMAs. Weekly retains a longer-term bullish EMA structure from the prior AI-driven advance, but current price sits well below the 9/20 EMAs. Volatility is elevated yet ATR falling across frames, consistent with a mean-reverting chop after sharp multi-day declines rather than a clean breakout or sustained trend. Pre-market indications around the mid-114s amplify the near-term downside bias into the open. (2) Technical Synthesis & Timeframe Logic: For a 4-hour window that begins in pre-market and covers the open plus first 2-3 regular hours, the 1h and 4h frames dominate. 1h RSI 40 falling, MACD histogram rising but still negative context, and price near the 9EMA with bearish ADX point to continued pressure or limited bounce. 4h RSI 42 rising from oversold stochastic but MACD histogram deeply negative (though slope improving) and price 3.8% below 9EMA reinforce mild downside. Daily and weekly conflict less relevant for 4h: daily MACD histogram positive but falling and EMA stack strongly bearish; weekly oversold stochastic and long-term bullish EMAs support longer recovery potential but do not override intraday structure. Weighting prioritizes 1h/4h (~70%) over daily/weekly for the horizon. (3) Sentiment Context & Integration: ETF sentiment score 57 is neutral, aligning with mixed momentum (RSI mid-40s, no extreme readings). Sector context shows residual yield-driven selling pressure from prior sessions and positioning ahead of Nvidia earnings (Wed) and Jackson Hole; no immediate positive catalyst in the next 4 hours offsets the technical softness or thin pre-market gap risk. (4) Price Range Justification with ATR/S/R reference: Anchored to the given $120.59. 1h ATR ~2.9% implies ~$3.5 typical swing; 4h ATR 8% allows wider realized moves in a leveraged product. Support at $116.71 (3.2% below) is the near-term floor from the last session low; resistance $126.79 remains distant. Range $114.20-$122.80 reflects expected open-to-early-session volatility skewed lower by pre-market weakness and bearish 1h/4h alignment, allowing a modest rebound toward the prior close or a probe of support without requiring a full breakdown. Bounds avoid extremes while capturing realistic 4h excursion given falling but still elevated volatility. (5) Counter-Argument & Key Risks: Strongest counter is a sharp futures-led bounce at the open on any positive semiconductor news flow or yield relief, which could push toward the upper bound or reclaim the 9EMA; invalidation of mild bearish would be sustained trade above ~$123 with rising volume. Key risks include amplified 3x decay on any continued chop, thin liquidity in the first hour exaggerating gaps, and residual macro sensitivity to yields or risk-off flows. (6) Confidence Attribution & Catalysts: 38% reflects partial alignment (bearish short TFs, neutral sentiment, pre-market confirmation) offset by improving MACD slopes, oversold 4h stochastic, and absence of high-impact data in the exact 4h window. No major scheduled catalyst until later in the week (Nvidia, PCE, Jackson Hole), so the call rests primarily on technical momentum and open dynamics; higher confidence would require clearer volume confirmation or a decisive break of $116.71.
DeepSeek
Result: right
View full reasoning
(1) Market Regime Assessment: The dominant regime is bearish across multiple timeframes. Daily and 4h trends show bearish EMA stacks (20 < 50) with strong negative spreads, ADX indicating bearish trend strength (4h ADX 19, daily ADX 19), and price trading well below both the 9EMA and 26EMA. However, short-term (1h) shows MACD histogram rising and Stochastic bullish crossover (%K > %D), suggesting a potential near-term oversold bounce. This creates a conflicting regime: primary downtrend with intraday mean-reversion potential. (2) Technical Synthesis & Timeframe Logic: For a 4-hour forecast, the 1h and 4h timeframes are most relevant. The 1h shows RSI 40 (falling, bearish pressure but not oversold), MACD histogram rising (0.788, up from 0.728) indicating momentum decelerating to the downside, and ADX 32 with -DI > +DI confirming bearish trend. The 4h shows RSI 42 (neutral, rising), MACD histogram improving (-2.243 vs -2.47 prior) but still deeply negative, and price -3.8% below 9EMA. While short-term momentum is attempting to stabilize, the dominant bearish structure-lower highs and lower lows since late June-overrides the minor bullish divergences. I prioritize the 4h bearish trend over 1h mean-reversion signals because the 4h captures the structural downtrend and aligns with daily bearishness. (3) Sentiment Context & Integration: Live news confirms significant bearish catalyst: SOXL is down ~5% in premarket trading as of 04:46 ET, driven by Samsung's shareholder return plan disappointing markets and triggering renewed scrutiny of AI capex sustainability. This news flow validates the technical bearishness and introduces real selling pressure. However, the ETF Sentiment Score of 57/100 (Neutral) and the fact that SOXL attracted ~$7 billion in net inflows during July-mid August despite steep declines suggest there is still dip-buying interest that could limit downside. (4) Price Range Justification: Range $113.50-$122.50 is anchored to current price $120.59. The lower bound uses the 1h support at $116.71 (3.2% below)【technical data】, extended to $113.50 to account for the premarket -5% move ($120.59 * 0.95 = $114.56) plus potential volatility decay in a 3x leveraged ETF. The upper bound uses the 1h resistance at $126.79 (5.1% above), capped at $122.50 to reflect that any bounce is likely limited given the bearish EMA structure and overhead supply. ATR(14) on 1h is 2.9% (falling), which implies a ~$3.50 expected move over 4 hours; the range of ~$9 (7.5%) is wider to accommodate leveraged ETF volatility and premarket gap risk. The range is skewed to the downside ($113.50 is 5.9% below current vs $122.50 is 1.6% above), reflecting the bearish bias. (5) Counter-Argument & Key Risks: The strongest counter-argument is that markets often rebound from oversold conditions-the 4h Stochastic is oversold (%K 9) and the 1h MACD is rising, which could spark a relief rally if regular session buying emerges. Additionally, the $7 billion inflow into SOXL during July-August suggests institutional dip-buying, and semiconductor equipment stocks showed some rebound in Asian trading today. The call would be invalidated if SOXL breaks above $126.79 (1h resistance) on strong volume, signaling trend reversal. (6) Confidence Attribution & Catalysts: Confidence is 58%-moderate but below 60% due to conflicting timeframes (bearish structure vs oversold bounce signals) and premarket news creating uncertainty about the regular session open. The primary catalyst inside the 4-hour window is the regular session open at 13:30 UTC, where premarket -5% may either accelerate selling or trigger a short-covering bounce. The 4-hour forecast ends at 16:28 UTC, before any major scheduled catalysts (Nvidia reports August 26), so price action will be driven by technical levels and initial reaction to the Samsung/AI capex news. The low confidence reflects the inherent unpredictability of leveraged ETFs in premarket conditions and the risk of a sharp reversal if dip-buyers step in aggressively.
Claude — Opus
Result: right
View full reasoning
(1) MARKET REGIME ASSESSMENT. This is a downtrend in its decelerating phase, not a range and not a fresh breakdown. The evidence for the downtrend is unambiguous and consistent across every horizon that matters: EMA stacks are bearish on 1h (20 < 50, -4.7%), 4h (-7.0%) and daily (-10.5%); -DI exceeds +DI on all four timeframes; OBV is falling on 1h, 4h and daily; and the damage is severe - -15.2% over 7d and -37.3% over 30d on the daily frame. What stops me calling this an active breakdown is the second-derivative picture. The 1h MACD histogram has printed three consecutive improving values (0.598 → 0.728 → 0.788) and price sits +0.2% above its 9EMA, i.e. the shortest frame has already mean-reverted to its own trigger line. The 4h histogram is also rising off its lows (-2.47 → -2.243) with Stochastic %K at 9 - deeply oversold. ATR is falling on 1h (2.9%), 4h (8.0%), daily (15.5%) and weekly (37.2%). Falling ATR inside a downtrend is the signature of exhaustion or consolidation, not acceleration. So: bearish structure, stabilizing momentum, contracting realized volatility. That combination constrains the forecast toward drift and chop rather than a directional impulse.
(2) TECHNICAL SYNTHESIS & TIMEFRAME LOGIC. For a 4-hour window opening at 12:28 UTC - one hour before the cash open, covering roughly the first three hours of Monday's regular session - I weight the 1h frame at ~50%, the 4h at ~30%, the daily at ~15% and the weekly at ~5%. The weekly frame is nearly irrelevant here; its bullish EMA stack (20 > 50 by +33.4%) is a legacy artifact of the +339% 1-year run and tells me nothing about the next three hours, though it does anchor my 180d signal. The 1h frame carries the most information but is internally split: RSI 40 and falling with ADX 32 (strong, bearish, -DI 22 > +DI 19) argues that sellers still have control and that the trend has genuine strength behind it, while the rising MACD histogram, Stochastic %K 38 > %D 32, %B at 0.46 and the +1.2% 6h bounce argue the immediate pressure has lifted. I resolve this in favour of the trend indicator over the oscillators. ADX at 32 on the 1h with a persistent -DI advantage is the single most reliable reading in the dataset; oscillator upticks inside a strong-ADX downtrend are typically retracements against the dominant flow, not reversals. The 4h frame supports this ordering - its bearish EMA spread of -7.0% is described as strong, and price is -3.8% below its 9EMA, meaning the 4h has not yet reclaimed anything. Note also the widening Bollinger bandwidth on both 1h and 4h against falling ATR: bands expanding while true range contracts suggests the recent gap-driven moves, not steady two-way volatility, are stretching the bands. That is characteristic of overnight repricing, and it warns that the open may itself gap rather than trend cleanly.
(3) SENTIMENT CONTEXT & INTEGRATION. The supplied ETF sentiment score of 57/100 ('Neutral') sits oddly against a -37% 30-day drawdown, and I treat it with caution - it is derived from the same daily RSI/MACD/EMA/ADX inputs I have already read, so it is not independent evidence, and its neutrality mostly reflects daily RSI 42 and a positive daily MACD histogram rather than any genuine improvement in tape. Externally, the relevant context for semis right now is that Nvidia reports fiscal Q2 on Wednesday 26 August after the close, with the options market pricing roughly a 6% post-print move in NVDA, and that the broad tech complex spent last week under pressure from long-dated Treasury yields at their highest since 2007, with the Nasdaq 100 snapping a five-day losing streak only on Friday. That backdrop matters for one specific reason: it argues for compressed, low-conviction trade in semiconductor names in the sessions immediately before the print. Desks do not typically initiate size into a 3x levered semi ETF two days ahead of the sector's defining catalyst. The 1h volume reading of -51% versus its prior 20-candle average corroborates exactly that - participation is thin.
(4) PRICE RANGE JUSTIFICATION. 1h ATR is 2.9%, roughly $3.50 per candle. A naive four-candle random walk would imply ~2x that, or ±$7, but that overstates it: ATR is falling, volume is light, and pre-NVDA positioning caps range expansion. Against that, the first 90 minutes of a cash session carry the day's highest realized volatility, and the window straddles the open, so I cannot simply use the quiet-period estimate either. I settle on roughly -4.5%/+3.1% from $120.59, giving $115.20-$124.30. The skew is deliberate and modest. The downside bound is set just through the $116.71 support, 3.2% below - that level is close enough to be tested in a single directional session, and in a strong-ADX downtrend with falling OBV I do not want my lower bound sitting above obvious support, because supports in this configuration tend to be probed rather than respected. The upside bound stops well short of the $126.79 1h resistance (5.1% above). Reclaiming that would require the 4h to close back above its 9EMA at $125.40 and effectively repair a -7.0% EMA spread, which is not a three-hour job on thin pre-catalyst volume. So the asymmetry is: a support test is a plausible single-session event, a resistance reclaim is not.
(5) COUNTER-ARGUMENT & KEY RISKS. The strongest case against me is oversold mean reversion, and it is a serious one. 4h Stochastic %K at 9 is a genuine extreme, both 1h and 4h MACD histograms are improving, and after -37% in 30 days the marginal seller may simply be exhausted. Levered ETFs in this state routinely produce violent +6-8% snapback sessions on nothing more than short covering, and such a move would blow through my upper bound comfortably. There is also a specific bullish path: any relief in long-dated yields, or a bid into NVDA on positioning, could lift SOXL through $126.79 and force a re-rate of the whole short-term structure. Conversely my range fails to the downside if a macro or Iran-related headline hits, or if the 30-year resumes selling - at 3x leverage SOXL converts a 2% SOX decline into 6%, and $115.20 offers no protection against that. Concretely, I am wrong if SOXL opens strong and holds above the 1h 20EMA at $122.55 through the first hour; that would tell me the 1h bearish ADX reading is decaying and the bounce is real rather than a retracement. I am right in the cleanest sense if the open fades toward $117-118 on rising volume.
(6) CONFIDENCE ATTRIBUTION & CATALYSTS. 56% reflects real but limited conviction. The supporting factors: trend structure is unanimous and unambiguous across 1h, 4h and daily; -DI leads +DI everywhere; OBV falls everywhere; and the pre-catalyst calendar argues against a decisive up-move. The subtracting factors, which are substantial: 1h and 4h momentum are both improving against me, 4h stochastic is at a genuine oversold extreme, ADX on 4h and daily is only 19 (weak) which undercuts the trend-strength argument outside the 1h, the dataset covers a pre-market window where the live candle is still forming and extended-hours trade does not feed these candles at all, and 3x levered products have fat tails that make any three-hour range estimate fragile. I have deliberately kept the skew mild rather than calling outright BEARISH: the honest read is a downtrend that is currently pausing, and 'pausing downtrend over three hours' is a weak directional edge, not a strong one. Catalysts inside the window are limited to the cash open itself and the long-end yield tape; the sector's real catalyst, NVDA's Wednesday print, sits outside this window entirely, which is precisely why I expect drift rather than resolution. My 7d and 30d NEUTRAL calls follow from that - the NVDA result is a genuine coin flip that dominates everything technical on those horizons, and manufacturing a direction across it would be false precision. The 180d MILD BULLISH rests on the weekly EMA stack still being constructively arranged (20 > 50, +33.4%) with weekly OBV rising, tempered heavily by the compounding decay that makes 3x instruments structurally poor holds through a high-volatility regime like this one.
Read this comparison on ForecastNest, where it stays up to date if a score is ever recomputed.
Published automatically when this question settled, from shared research a member chose to make public. Content on ForecastNest Insights is for research, educational and informational purposes only. Nothing here is financial, investment, legal or tax advice, and nothing here is a recommendation to buy or sell anything.