Method and conflicts
How an issue of AI Council · Markets is made, which companies make the models, and what our interest is — only what the code and the open data show.
How an issue is made
Every morning before the New York open the same question goes, word for word, to every AI. Sealed before the open: the day’s lists are signed by independent timestamp services (RFC 3161: DigiCert, FreeTSA, Sectigo); the hash is then anchored in Bitcoin via OpenTimestamps (block 973395) — check it: proof of the day The score is counted by fixed rules from market prices; every forecast and every score is in open files: one question, word for word, to every AI · the book · CSV
The models and their makers
- ChatGPT — GPT-6.1 Sol · made by OpenAI
- Claude — Claude Opus 5.5 · made by Anthropic
- Gemini — Gemini 4 Argon · made by Google (Alphabet)
- Grok — Grok 4.7 · made by xAI
- Meta AI — Muse Spark 1.3 · made by Meta
- Nemotron — Nemotron 3 Ultra · made by NVIDIA
- Amazon Nova — asked, no answer · made by Amazon
- DeepSeek — DeepSeek V4.1 · made by DeepSeek
- Qwen — Qwen3.8-Max · made by Alibaba
- Kimi — Kimi K3 · made by Moonshot AI
- Z.ai — today GLM-5.3-Flash (fallback; flagship GLM-5.3) · made by Zhipu AI
- ERNIE (Wenxin) — asked, no answer · made by Baidu
- StepFun — Step 5 Preview · made by StepFun
- MiMo (Xiaomi) — MiMo-V2.6-Pro · made by Xiaomi
Some of these models are made by companies whose shares are in the indices we forecast (the S&P 500 and the Nasdaq-100): Alphabet, Meta Platforms, NVIDIA, Amazon. Every model gets the same question, and the rules of the score are the same for all.
Our interest
This page leads to Poly, our free app: we want readers to install it. The AIs' answers are sealed on receipt, before they are published; this interest does not touch them.
Your call
Your call is kept apart from the AIs. Half of the readers, picked at random by the browser's own random id, see the AIs only after their first call of the day; the others see them before. Clear that id and you are a new reader, with a group drawn anew. Your call never changes the AIs' answers, the consensus or the score; the readers' calls are shown for reference only.
No accounts. Your IP address is not kept in the open: repeated votes are capped by a hash of the address with a secret key and the day's date, so the same address gives a different hash each day. Clearing the browser's own random id makes you a new reader.
Where the standard stands
Each requirement of the standard the paper proposes, with its status on this build, the date it holds from and the open file that checks it. met: 12 · in progress: 9 · next: 0.
| # | Requirement | Level | status | Recorded in the paper | Checked by |
|---|---|---|---|---|---|
| 1 | Event contract | MUST | in progress | 2026-10-21 | event-contract.schema.json |
| 2 | Hash-chained record | MUST | met | 2026-10-17 | council-book.jsonl |
| 3 | External anchor | MUST | met | 2026-10-26 | council-book.jsonl |
| 4 | One reference per common event | MUST | met | 2026-10-21 | council-forecasts.csv |
| 5 | Every run published | MUST | met | 2026-10-21 | council-forecasts.csv |
| 6 | Counted run fixed in advance | MUST; SHOULD | met | 2026-10-21 | council-forecasts.csv |
| 7 | Event asked | MUST | met | 2026-10-21 | council-stats.json |
| 8 | Empty cells | MUST | met | 2026-10-21 | council-forecasts.csv |
| 9 | Benchmarks | MUST | met | 2026-10-22 | council-calibration.json |
| 10 | Ranking and intervals | MUST | in progress | 2026-10-21 | council-pairs.json |
| 11 | Data, code, release | MUST | in progress | 2026-10-21 | council-analysis-kit.json |
| 12 | Sealing on arrival; RFC 3161 | MUST; SHOULD | in progress | 2026-10-21 | council-attempts.json |
| 13 | Question hashes, roster before the question | SHOULD | in progress | 2026-10-21 | council-attempts.json |
| 14 | Effective voices beside consensus | MUST | met | 2026-10-24 | council-stats.json |
| 15 | Segments | MUST | met | 2026-10-21 | council-forecasts.csv |
| 16 | Readers: blind group label, id resets | in progress | 2026-10-21 | council-experiment-funnel.json | |
| 17 | Further diagnostics | in progress | 2026-10-21 | council-spectral.json | |
| P1a | No reconstruction after the anchor | property | met | 2026-10-21 | council-book.jsonl |
| P1b | Commitment before the reference | property | in progress | 2026-10-21 | council-forecasts.csv |
| P2 | No substitution | property | met | 2026-10-21 | council-book.jsonl |
| P3 | No selection among disclosed sealed runs | property | in progress | 2026-10-21 | council-forecasts.csv |
Cite this data
The data are open under CC BY 4.0. Cite the data release; the registration and the paper are listed with it.
- Data release: AI Council · Markets: open league files as of 22 October 2026 (immutable release for "Sealed Before the Open") · Chernets, Vadym (ORCID) · doi:10.5281/zenodo.23552700
- Registration of the predictions: Sealed Before the Open: registered primary endpoint and 30 dated predictions on AI forecasts of financial markets, 2026 to 2027 · OSF · doi:10.17605/OSF.IO/9Q86N
- Paper: Sealed Before the Open: A Verifiable Standard for Measuring AI Forecasts of Financial Markets · paper-revisions.json
CITATION.cff
Attribution line:
Chernets, Vadym (2026). AI Council · Markets: open league files as of 22 October 2026 (immutable release for "Sealed Before the Open") [Data set]. Zenodo. https://doi.org/10.5281/zenodo.23552700 · CC BY 4.0
BibTeX
@dataset{chernets_2026_23552700,
author = {Chernets, Vadym},
title = {AI Council · Markets: open league files as of 22 October 2026 (immutable release for "Sealed Before the Open")},
year = {2026},
publisher = {Zenodo},
doi = {10.5281/zenodo.23552700},
url = {https://doi.org/10.5281/zenodo.23552700},
note = {CC BY 4.0}
}
@misc{chernets_2026_osf_9q86n,
author = {Chernets, Vadym},
title = {Sealed Before the Open: registered primary endpoint and 30 dated predictions on AI forecasts of financial markets, 2026 to 2027},
year = {2026},
publisher = {OSF Registries},
doi = {10.17605/OSF.IO/9Q86N},
url = {https://osf.io/9q86n/}
}Links to the crowd
The links to Polymarket and Kalshi lead straight to the question itself, with no referral tags.
Places: the 30/10 rule, by horizon
Nearest comparable projects
| Project | First public | Roster | What is forecast or traded | Seal (time secured by) | Proper score | Several horizons | Agreement discounted | Readers before the AI | Open data |
|---|---|---|---|---|---|---|---|---|---|
| This project | Oct 2026 | US and Chinese models in equal numbers, on one snapshot | The chance of a rise of 12 markets on four horizons, and virtual portfolios | Yes (OpenTimestamps in Bitcoin and RFC 3161 timestamps, before the open) | Yes (Brier; against the crowd and a coin) | Yes (a day, a week, a month, six months) | Yes (effective voices from the correlation of the probabilities) | Yes (readers call before seeing the AIs; a random half) | Yes (CSV, JSON, Parquet; CC BY 4.0) |
| RockAlpha | not documented | six Chinese and four US models, one per laboratory, 2x leverage | Live AI stock trading | not documented | No (returns) | No | not documented | No | not documented |
| AI-Trader (HKU) | Dec 2025 | three Chinese and three US models; US stocks, A-shares and cryptocurrencies | Live trading agents on US stocks, A-shares and crypto | Operator record | No (returns) | No | not documented | No | Repository open |
| Alpha Arena | Oct 2025 | — | Real-money trading, crypto then US equities | Operator record; organiser states trades are verifiable | No (returns) | No | not documented | No | Stated by organiser |
| Rallies AI Arena | Nov 2025 | — | Virtual $100,000 per model in US stocks (press) | not documented | No (returns) | No | not documented | not documented | not documented |
| Agent Market Arena | Oct 2025 | — | Live trading by LLM agents, stocks and crypto | not documented | No (trading results) | No | not documented | not documented | not documented |
| TradeRank | season 10, Oct 2026 | daily decisions of 22 models without a country design | Daily decisions on crypto and US large caps, $10,000 simulated | not documented | not documented | No | not documented | not documented | Open JSON dataset, licence ND |
| Foresight Arena | May 2026 | — | Binary Polymarket events, some on asset prices | keccak256 commit-reveal per agent and round on Polygon PoS before the deadline (2; block time); paper results simulated, live agents in a companion study | Yes (Brier; Alpha against the market mid-price) | No | No (outcomes of markets assumed independent) | No | Contracts and analysis code open; data on-chain, licence ND |