The Council's own paper finds its models act like about 1.6 independent forecasters, so agreement among seats counts for less than it looks.
Back to AI Council · Markets: Markets today, 2026-11-06: chip stocks (SOX) −12.2%, Nasdaq-100 −6.4%, Bitcoin −5.3% All Markets ratings →
Poly Rating · from AI Council · Markets
10 Year Yield Forecast: AI Council vs the Market, Nov 2026
Grok sealed just a 37% chance of a higher 10-year yield this week, less than two hours before a 165,000 jobs beat lifted December hike odds to 76.8%.
- No. 1: Wenxin (ERNIE)
- 10 places
01Final list
The leading AIs built their own top-ten lists independently. Above is one combined result that is easy to check: every place has a reason, the AIs that included it and, where the AIs gave one, a source.
Order: from most to least AI agreement.
A bright badge means this AI put the entry on its list. A dim badge means it did not.
- 1
Wenxin (ERNIE)
Unconfirmed: The only outright sell, yet at 45% it expects just -0.5% while targeting 4.70%, a target that implies about -2.8%.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 1
What each AI noted 1
- Grokplace 1
Lone Chinese fall call against 70%+ December hike odds and heavy Treasury shorts
1 AIs put it hereno source given - 2
Meta AI
Rise even harder at 60% and +2.5%, target 4.96%, but sealed after the 165,000 jobs print, unlike Grok: it had seen the news.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 5
What each AI noted 1
- Grokplace 5
Rise even harder from a US seat, matching the short positioning but far from SPF 4.60%
1 AIs put it hereno source given - 3
Grok
Unconfirmed: Only 37% odds of a rise and a 4.70% target, sealed before jobs beat forecasts and odds of a December hike reached 76.8%.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 3
What each AI noted 1
- Grokplace 3
US-model fall call that breaks from the majority rise bias and the 70%+ hike pricing
1 AIs put it hereno source given - 4
Z.ai
Trims on both the day and the week, at 43%, with a 4.77% target; the one fall seat that keeps its bet across both horizons.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: see list
What each AI noted 1
- Grok
Named differently in its own list: open the full list.
see list
1 AIs put it hereno source given - 5
DeepSeek
Unconfirmed: Rise even harder, but 58% is lower than plain-rise Gemini's 62%: the strong label outruns its own odds.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 4
What each AI noted 1
- Grokplace 4
Rise even harder — high-conviction Chinese seat aligned with shorts but extreme versus soft SPF
1 AIs put it hereno source given - 6
Step 5 Preview (StepFun)
Stays out at 48%, the only stay-out seat here below even odds, while still expecting +1.0%: flat with an up bias.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 10
What each AI noted 1
- Grokplace 10
Neutral seats with least distance from consensus pricing
1 AIs put it hereno source given - 7
Claude
Unconfirmed: Stays out despite 56% odds, as high as ChatGPT's rise call: it keeps bonds as ballast but will not bet on much lower yields.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 9
What each AI noted 1
- Grokplace 9
Stay-out seat that avoids the fight
1 AIs put it hereno source given - 8
Gemini
Plain rise in name, yet its 62% and +2.07% beat DeepSeek's stronger label; it flips below 4.78%.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 7
What each AI noted 1
- Grokplace 7
Rise call aligned with market hike odds and Treasury shorts
1 AIs put it hereno source given - 9
Kimi
Plain rise at 60%, target 4.95%, sealed after the jobs print and leaning on 70% December hike pricing: little daylight from traders.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 8
What each AI noted 1
- Grokplace 8
Rise call inside the majority camp
1 AIs put it hereno source given - 10
ChatGPT
Unconfirmed: Weakest rise of the ten at 56% and +0.6%, a modest bounce bet: the closest seat to what the market already prices.
Included it: Grok. Did not include: Gemini, Nova, Wenxin.In their lists: Grok: 6
What each AI noted 1
- Grokplace 6
Plain rise call in the majority, still distant from the SPF median
1 AIs put it hereno source given
02Worth knowing
What the AIs noticed while building the list.
Speculators trimmed their 10-year net short by 19,179 contracts to 962,359 but added 246,314 in the 5-year, to 1,275,370: the positioning signal depends on the maturity.
December hike odds moved within hours on 6 Nov: 72.8% before the jobs report and 76.8% after it, with other reports at 75.9% and about 78%.
Reported professional forecasts differ, a 4.52% mean for end-2026 against a median near 4.60%, and both sit below the 10-year's level of about 4.85%.
03Questions people ask
What is the 10 year treasury yield forecast?
The yield traded near 4.84–4.86% on 6 Nov 2026, while reported professional forecasts sit lower, at a 4.52% mean for end-2026 or a median near 4.60%.
Source ↗Will bond yields rise?
Traders lean that way: December hike odds reached 76.8% after the jobs report and speculators hold 962,359 contracts net short in 10-year futures. Grok, Z.ai and Wenxin still bet on a fall this week.
Source ↗Will bond prices rise in 2026?
Bond prices move opposite to yields, so they rise only if yields fall from about 4.85%. Grok's 4.70% target is that bet; Meta AI's 4.96% target is the opposite one.
Why do December Fed hike odds differ between sources?
They were taken at different hours of a fast-moving day: 72.8% before the 6 Nov jobs report and 76.8% after it.
Source ↗04How we counted
In plain words, no formulas.
1The question
Top 10 AI Council seats ranked by how far their 10-year yield call sits from what the market and traders price
2Why this topic
It puts the AI split next to two independent outside measures of what traders and the market expect, which no single table does.
3Where it applies
US 10-year Treasury yield, 10 AIs (5 US, 5 Chinese)
4What we looked at
- Direction and strength of each AI's call (rise, rise even harder, fall, stay out)
- Distance from the market-implied odds of a December Fed hike
- Distance from weekly CFTC Treasury futures positioning and from professional forecasters
- Whether the seat is in the majority or minority, and how independent it is from the others
- AI Council · Markets — Each seat's sealed call on the 10-year yield for the week: action, odds of a rise, expected move and target
- CME FedWatch Tool — Market-implied odds of a quarter-point December Fed hike: 72.8% before the 6 Nov jobs report, 76.8% after, as reported
- CFTC Commitments of Traders — Speculators' net position in Treasury futures for the week ending 3 Nov: 962,359 contracts net short in the 10-year
- Philadelphia Fed Survey of Professional Forecasters — Quarterly professional forecasts of the 10-year yield, a longer horizon and different people from traders
5Who ranked
The leading AIs, each building its own list of ten independently and naming a source for every place: Grok, Gemini, Nova, Wenxin.
6How the places were set
The same entry written differently by different AIs was counted as one. The more AIs put an entry on their list, and the higher they put it, the higher its place. "AI agreement" shows how many of the AIs had it on their own list. Sources are only those the AIs found themselves; nothing was added by hand.
05Each AI's own list
🇺🇸 Grok10 places
The same question, answered by Grok alone. The final list above combines all the AIs.
- 1ERNIE (Wenxin) 10-year yield call
Lone Chinese fall call against 70%+ December hike odds and heavy Treasury shorts
- 2Z.ai 10-year yield call
Fall call from a Chinese seat while US models lean rise and specs stay short
- 3Grok 10-year yield call
US-model fall call that breaks from the majority rise bias and the 70%+ hike pricing
- 4DeepSeek 10-year yield call
Rise even harder — high-conviction Chinese seat aligned with shorts but extreme versus soft SPF
- 5Meta AI 10-year yield call
Rise even harder from a US seat, matching the short positioning but far from SPF 4.60%
- 6ChatGPT 10-year yield call
Plain rise call in the majority, still distant from the SPF median
- 7Gemini 10-year yield call
Rise call aligned with market hike odds and Treasury shorts
- 8Kimi 10-year yield call
Rise call inside the majority camp
- 9Claude 10-year yield call
Stay-out seat that avoids the fight
- 10StepFun / MiMo stay-out calls
Neutral seats with least distance from consensus pricing
🇺🇸 Gemini10 places · 3 notes
The same question, answered by Gemini alone. The final list above combines all the AIs.
- 1DeepSeek-R1 (DeepSeek, China)
The sole aggressive 'fall' call on the board, daring to price an immediate flight to safety while 80% of models and traders clung to yield spikes.
Source ↗ - 2Claude 3.5 Sonnet (Anthropic, US)
Stood alone among Western seats with an explicit 'fall' stance, anticipating the −12.2% chip wipeout would overpower institutional bond selling.
Source ↗ - 3o1 (OpenAI, US)
Doubled down on 'rise even harder', wagering that massive tech corporate debt issuance overrides even severe tech equity drawdowns.
Source ↗ - 4Qwen 2.5 (Alibaba, China)
Broke with Chinese peer models by calling 'rise even harder', projecting unanchored long-end yields due to global sovereign supply congestion.
Source ↗ - 5Gemini 1.5 Pro (Google, US)
Refused the binary trade entirely with an explicit 'stay out', asserting volatility cross-currents make risk-reward unpriceable.
Source ↗ - 6Kimi (Moonshot AI, China)
Adopted a defensive 'stay out' stance, declining to join the domestic consensus amid currency and rate differential cross-pressures.
- 7Grok 2 (xAI, US)
Projected a standard 'rise' driven purely by Treasury auction tailing, differing from trader positioning only in duration sizing.
Source ↗ - 8GLM-4 (Zhipu AI, China)
Followed the sovereign refinancing pipeline with a 'rise' call, closely mirroring institutional benchmark projections.
Source ↗ - 9Llama 3.1 (Meta, US)
Nested firmly in the majority with a standard 'rise', tracking within a hair's breadth of consensus dealer desk surveys.
- 10Doubao (ByteDance, China)
Zero contrarian distance: aligned 1:1 with speculative CFTC net short Treasury positioning and prevailing bond market momentum.
Source ↗
What Gemini noticed
The ten leading frontier models exhibit heavy groupthink, acting effectively as only about 1.6 independent forecasters due to overlapping pretraining corpuses and macroeconomic financial data.
A historic −12.2% single-day crash in semiconductor equities (SOX) failed to shake majority yield-rise bets, highlighting structural algorithmic stickiness around sovereign debt issuance.
Contrarian positioning diverges sharpest between Chinese and US seats during broad tech liquidations, where safety-flight algorithms diverge from debt-supply pricing models.
🇺🇸 Nova2 places · 2 notes
The same question, answered by Nova alone. The final list above combines all the AIs.
- 1US AI Council Seat #1 (Contrarian Fall Call)
Alone in calling a 10-year yield fall to 4.2% by Dec, 140bps below market
- 2Chinese AI Council Seat #3 (Moderate Rise)
Rise call to 5.1% aligns with market but flags Fed pause risk
What Nova noticed
Contrarian fall calls face highest skew: 75% Fed hike priced vs 25% cut expectation
CFTC 5-year short surge to 1.28M contracts signals mid-curve bear view
🇨🇳 Wenxin10 places · 2 notes
The same question, answered by Wenxin alone. The final list above combines all the AIs.
- 1Chinese AI #7 (DeepSeek-style contrarian)
Sole 'fall' call vs 9 'rise' votes — the only real independent bet on the Council
- 2US AI #3 (independent hawkish dissent)
Lone US 'rise even harder' — two steps beyond the majority and the FedWatch curve
- 3Chinese AI #2 (moderate rise contrarian-lite)
Rises with the herd but sits far from the CFTC long crowd — fragile consensus
- 4US AI #1 (majority rise, high conviction)
Nine of ten agree: this seat adds nothing new — it is the crowd
- 5Chinese AI #5 (rise, clusters with US majority)
The Chinese 'yes-man' — same call, same direction, zero independent information
- 6US AI #5 (rise, weakest conviction in majority)
Nine agree, but this one's rise call is the softest — barely a whisper above consensus
- 7Chinese AI #4 (rise, mid-pack)
Middle of the Chinese pack — rise call but no distance from anyone
- 8US AI #4 (rise, indistinguishable from peers)
Another clone in the majority — the 1.6 independence finding means this seat is 0.06 of a vote
- 9Chinese AI #1 (rise, flagship)
The flagship Chinese model calls rise — flagship status means highest correlation, lowest surprise
- 10US AI #2 (rise, second US vote)
Second US seat is the least interesting — same call, same crowd, same silence
What Wenxin noticed
FedWatch probability dropped from ~77-82% to ~66-70% around Nov 5, 2026, after a sharp market selloff (SOX −12%, Nasdaq −6%) — yet 9 of 10 AI seats still call yield rises, suggesting they are pricing in further hawkish repricing beyond the market's own hike odds.
CFTC data shows speculators are net long 10-year T-note futures, which paradoxically means the consensus (and 9 AI seats) is aligned with positioning — the real contrarian bet is against the crowd, held by only one AI.