The NBER declares that a US recession began during 2026.
Persistent inflation and interest rate policy have driven speculation over whether the US economy enters an official recession before the end of 2026.
Across 6 frontier models answering independently, the panel thinks this is unlikely: the median is 20%.
They broadly agree, 27 points between GPT at 39% and DeepSeek at 12%.
Nothing is settled until 10 Oct 2027, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GPTOpenAI
It turns on whether weakening jobs and real income become broad, sustained contraction rather than a soft landing.
- The NBER Business Cycle Dating Committee must later identify a peak in 2026; it often dates recessions months after they begin.
- Since 1945, NBER recessions have occurred roughly once per 6–7 years, but expansions can persist through restrictive policy.
- A sharp rise in unemployment toward 5% or a multi-month payroll decline would materially raise the odds of an NBER peak in 2026.
39% -
GeminiGoogle
Turns on whether an unannounced downturn already began in 2026 or starts in Q4 and is recognized swiftly by NBER.
- NBER takes an average of 12 months to date peaks, leaving little runway before October 2027 for a late-2026 onset to be called.
- Unconditional annual recession risk is roughly 15%, and ten months of 2026 have passed without an evident widespread contraction.
- A sharp spike in Q4 2026 unemployment triggering the Sahm rule followed by rapid NBER designation would flip this to likely.
22% -
QwenAlibaba
NBER typically declares recessions 6-12 months after they begin; a 2026 start needs clear deterioration by mid-2027 at latest.
- NBER declared the 2020 recession in June 2020 (fast due to severity) but took until Dec 2001 for the 2001 recession—11 months after it ended.
- For a declaration by Oct 2027, the recession would need to have started by roughly mid-2026 with unmistakable signals (employment, income, sales) by early 2027.
- Current labor market and GDP data through late 2025 showed resilience; a sharp enough turn to trigger NBER within 12 months requires a shock-level event, which is…
22% -
GrokxAI
NBER must both see a 2026 peak and announce it before Oct 2027.
- NBER typically dates peaks 8–15 months after they occur; a late-2026 start would miss the 2027-10-10 cutoff.
- Recessions occur ~once per 6 years; the 2020 episode and subsequent expansion make a 2026 peak below that base rate.
- A sharp Q4-2026 GDP drop plus two consecutive NFP declines would raise odds; continued 2%+ real growth would kill them.
18% -
ClaudeAnthropic
Hinges on whether labor market and GDP data deteriorate sharply enough in 2026 to trigger an NBER-dated downturn.
- NBER recession calls require broad declines in employment, income, and output, which have no clear current trajectory without more data.
- Base rate for a given year containing an NBER-dated recession start is roughly 10-15% historically outside acute shocks.
- Elevated rates and inflation raise recession risk somewhat above baseline, but no major financial shock or labor collapse is yet evident as of Oct 2026.
16% -
DeepSeekDeepSeek
NBER dating lags badly, so this needs a 2026 peak that is already visible in data by mid-2027.
- NBER's Business Cycle Dating Committee typically announces peaks 9-21 months after they occur, so a 2026 peak would likely be declared in late 2027 or 2028.
- Real GDP grew through mid-2026 and unemployment stayed near 4.2%, so no broad contraction has yet appeared in the data the committee weighs.
- The committee requires a significant, pervasive decline in activity, not just two negative GDP quarters, and it has never dated a peak within the same calendar year it…
12%
6 of 6 models answered · 27 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 10 Oct 2026. Forecast by google/gemini-3.8-flash, anthropic/claude-sonnet-5, openai/gpt-5.6-terra, x-ai/grok-4.6, deepseek/deepseek-v4.1-flash, qwen/qwen3.8-max-0902 via OpenRouter.