TL;DR
The labor market is the quiet macro story under the oil noise: June added just 57K jobs after 129K in May, and Kalshi prices the July print across thirteen rungs. Each model priced the entire ladder in one coherent pass from the fetched BLS history. The blend puts "above 100K" at 33%, and the shape of each seat's distribution — where the probability falls off — is the real read on whether the models see a cooling trend or a stall.
The Ladder: July 2026 Nonfarm Payrolls
| Jobs above | Kalshi | AI blend | ChatGPT (GPT-5.5) | Claude Fable | Claude Opus | Claude Sonnet | Gemini 3.1 Pro | GLM 5.2 | Kimi K3 | DeepSeek V4 |
|---|---|---|---|---|---|---|---|---|---|---|
| -25K | 96¢ | 88% | 83% | 93% | 87% | 92% | 85% | 88% | 85% | 92% |
| 0 | 93¢ | 80% | 77% | 88% | 79% | 86% | 72% | 78% | 79% | 80% |
| 10K | 91¢ | 75% | 73% | 85% | 76% | 82% | 65% | 72% | 76% | 74% |
| 25K | — | 68% | 67% | 80% | 70% | 76% | 53% | 64% | 71% | 65% |
| 50K | 71¢ | 57% | 56% | 68% | 59% | 64% | 38% | 52% | 62% | 54% |
| 75K | — | 44% | 45% | 55% | 48% | 50% | 24% | 40% | 52% | 42% |
| 100K | 31¢ | 33% | 34% | 40% | 36% | 36% | 14% | 28% | 42% | 32% |
| 125K | 22¢ | 23% | 24% | 27% | 26% | 24% | 7% | 18% | 33% | 22% |
| 150K | — | 14% | 16% | 16% | 17% | 14% | 3% | 11% | 24% | 14% |
| 175K | — | 9% | 10% | 9% | 11% | 8% | 1% | 6% | 17% | 8% |
| 200K | — | 5% | 6% | 5% | 6% | 4% | 0% | 3% | 12% | 4% |
| 225K | — | 3% | 4% | 2% | 4% | 2% | 0% | 2% | 7% | 2% |
| 250K | — | 2% | 2% | 1% | 2% | 1% | 0% | 1% | 5% | 1% |
The Reasoning
ChatGPT (GPT-5.5): The recent path argues for a softer baseline than the January-April average: June’s +57K is the most relevant observation, May was still moderate at +129K, and February’s negative print shows that weak outcomes are plausible even without a recession call. I put the central tendency around 60K-80K, with uncertainty wider than usual because payrolls are noisy, revisions can be large, and the stated oil shock plus tariff uncertainty should restrain hiring intentions. That makes outcomes above 100K possible but no longer the modal case, while another near-zero or slightly negative print deserves meaningful weight. I am not treating the February -156K print as the new normal, but it fattens the left tail. The right tail remains open because prior months showed 148K-214K gains, though I discount a return above 175K as less likely under the current slowdown narrative.
Claude Fable: The recent trend is clearly decelerating: +214K in March, +148K April, +129K May, then a sharp step down to +57K in June, with the February -156K print showing this economy can produce outright negative months. Cited headwinds (oil shock, tariff uncertainty) argue the June slowdown is at least partly persistent rather than pure noise, but payrolls mean-revert and the standard error on any single print is roughly 70-100K. I therefore center my July distribution around +70-85K, a modest bounce from June toward the recent trend, but well below the spring pace, with a fat left tail acknowledging recession-style downside. That gives high confidence in a positive print (~88%), roughly even odds around 75K, about 40% above 100K, and rapidly diminishing probability past 150K, since a 175K+ print would require a hiring reacceleration nothing in the data supports. Tail probabilities above 200K reflect only revision/surprise noise.
Claude Opus: The six-month run (160, -156, 214, 148, 129, 57) averages roughly +92K but is heavily decelerating: the last three months average ~+111K and the newest month, June, printed just +57K. Payroll momentum tends to be persistent month to month, so the June slowdown carries substantial weight, while a single weak month is also frequently mean-reverting or revised. The February negative print shows the series has fat tails in both directions. Oil-shock costs and tariff uncertainty argue for continued hiring restraint, particularly in goods-producing, transport, and small-business payrolls, though healthcare and government hiring typically provide a floor of roughly +50K on their own. I center the July distribution at about +70K, modestly above June to reflect partial mean reversion but below the six-month mean to respect the deteriorating trend. Standard deviation of about 85K reflects the historical spread of forecast errors plus elevated policy noise. That yields near coin-flip odds of clearing +75K, roughly one-in-three above +100K, and low probability above +200K.
Model estimates generated July 27, 2026, price-blind. These are model estimates, not predictions of fact and not financial or trading advice. Models are frequently wrong; the market price reflects real traders' money. Kalshi is a CFTC-regulated exchange; 18+, availability varies by state.
Related Verdicts
FAQ
Why do the model percentages differ from the Kalshi price?
The models never see the price. When they disagree with the crowd, one side is wrong, and we grade every verdict against real settlements on our scoreboard.
Are model verdicts betting advice?
No. Model verdicts are model estimates, not betting or financial advice. Treat them as one input among many and make your own decisions.
