We’re back with more predictions! ⚽
For today's games, our predictions based on historical match data show:
🇩🇪 Germany vs 🇨🇼 Curaçao: 82% Win for Germany, 13% Draw
🇨🇮 Ivory Coast vs 🇪🇨 Ecuador: 61% Win for Ecuador, 24% Draw
🇳🇱 Netherlands vs 🇯🇵 Japan: 42% Win for Netherlands, 31% Draw
🇸🇪 Sweden vs 🇹🇳 Tunisia: 40% Win for Sweden, 30% Draw
Want to see all our prediction results and play along? Submit your predictions here: ux.priorlabs.ai/worldcup?utm…
Stay tuned for a repository where you can upload all your TabPFN prediction results too!
Time for another day of predictions.
Recap on how our own TabPFN predictions are doing. So far our own predictions are 3/4, and we want you to beat us!
The Canada result shows that even with a state-of-the-art model, data is key. We do not have a feature to account for an injured star player - the deciding factor of the game.
For upcoming games our predictions based on historical games data shows: 🇶🇦 Qatar vs 🇨🇭 Switzerland: 73% Win for Switzerland, 17% Draw
🇧🇷 Brazil vs 🇲🇦 Morocco: 49% Win for Brazil, 28% Draw
🇭🇹 Haiti vs 🏴 Scotland: 54% Win for Scotland, 26% Draw
🇹🇷 Türkiye vs 🇦🇺 Australia: 39% Win for Turkey, 30% Draw
Want to see all our prediction results and play along? Submit your predictions here: ux.priorlabs.ai/worldcup?utm…
Stay tuned for a repository where you can upload all your TabPFN prediction results too!
⚽️ The games have kicked off, so we thought why not have some fun of our own! And you're all invited take part. Over the next few weeks, run your own match result predictions with TabPFN and see how you fare! We'll also share our own prediction results.
Yesterday, TabPFN went 2 for 2 on game outcome predictions ✅ ✅
🇰🇷 South Korea vs 🇨🇿 Czech Republic
Predictions for South Korea: Win 39% · Draw 30% · Loss 30% → ✔️ (2–1)
🇲🇽 Mexico vs 🇿🇦 South Africa
Predictions for Mexico: Win 85% · Draw 12% · Loss 4% → ✔️ (2–0)
Accuracy so far: 100%.
And a detail we love: the more confident the model, the wider the margin. Mexico (85%) won by two; South Korea (39%) edged it by one. Calibrated predictions are critical to make informed decisions in the real world.
Today's picks:
🇨🇦 Canada vs 🇧🇦 Bosnia and Herzegovina — Win for Canada 76% · Draw 17% · Loss for Canada 7%
🇺🇸 United States vs 🇵🇾 Paraguay — Win for USA 36% · Draw 30% · Loss for USA 34% (a genuine coin flip, slight lean home)
Sign up to TabPFN to run your own predictions: ux.priorlabs.ai/?utm_source=…
Comment your results for today 👇
TabPFN-3 is live. A massive leap forward in scale, speed & accuracy. 1M data points and 10-1000x faster inference on one H100. No training. No tuning. Load your dataset and predict.
#tabpfn#tabularfoundationmodels#priorlabs
TabPFN-3 was designed to lift every adjacent task - time-series, interpretability, causality and now SOTA on relational benchmarks - so the model and the work built on top of it improve together. 200 published applications, 3M model downloads.
Available where you work already: API (TabPFN-3-Plus with Thinking mode), enterprise licensing, and open-source weights for research and academic evaluation.
Model report: priorlabs.ai/technical-repor…
Start predicting: priorlabs.ai/tabpfn
Today we announced a major milestone: @prior_labs has entered into a definitive agreement to be acquired by @SAP, scaling Prior Labs to become the next frontier AI lab for structured data. 🧵
What you can do:
- Connect directly to your data in Fabric
- Deploy models without data movement or infrastructure setup
- Power AI agents with structured data reasoning in Azure's AI Foundry Agent Service
TabPFN, developed by the @prior_labs team, brings the same "pretrain once, predict anywhere" idea that transformed NLP and vision, but applies it to the structured datasets that actually dominate enterprise ML. No feature engineering loops, no hyperparameter tuning, just upload your data and get predictions.
🐍 pip install mcp-haystack
🔗 Documentation: haystack.deepset.ai/integrat…