AI Models Prediction Market Data
Tick-by-tick trades and L2 order-book depth for 17 AI Models series on Kalshi & predict.fun — Year-end top LLM (KXLLM1), top-ranked AI, best coding model, Humanity’s Last Exam, AGI before date, Claude model release, capability spikes, OpenAI/Jony Ive hardware. 3,440 series-days covering March 2026 – present; 50.7K trades · $1.2M turnover over the last 30 captured days. Every contract captured co-located with nanosecond venue and receive timestamps — sold as day bundles, €1 per day, bought as a guest with instant download.
AI Models on Kalshi & predict.fun
Year-end top LLM, best AI coding model, top-ranked model on Chatbot Arena, Humanity’s Last Exam scores, AGI-by-date, model releases and OpenAI hardware.
Every series below is archived at full feed resolution — each contract file carries every trade print (price, size, side) and Level-2 order-book updates in one normalized, zstd-compressed schema. A series-day bundle holds all contracts of the series for one UTC day.
Most traded AI Models series (last 30 days)
Ranked by USD turnover from our own daily statistics — the series a market maker or bot builder should start with.
| # | Series | Venue | Trades (30d) | Turnover (30d) | Days | |
|---|---|---|---|---|---|---|
| 1 | Year-end top LLM | Kalshi | 40,323 | $1.0M | 30 | Buy days |
| 2 | AI capability spike | Kalshi | 718 | $104.7K | 30 | Buy days |
| 3 | Will any company achieve AGI before [date]? | Kalshi | 698 | $29.3K | 30 | Buy days |
| 4 | Top model | Kalshi | 1,192 | $25.9K | 30 | Buy days |
| 5 | AI Coding | Kalshi | 1,106 | $17.3K | 30 | Buy days |
| 6 | Best AI coding model | Kalshi | 1,151 | $13.2K | 30 | Buy days |
| 7 | Model hitting thresholds | Kalshi | 816 | $11.7K | 30 | Buy days |
| 8 | ZHIPU Up/Down 15m | predict.fun | 1,247 | $10.8K | 14 | Buy days |
| 9 | Humanity's Last Exam | Kalshi | 309 | $9.3K | 30 | Buy days |
| 10 | Claude Model Release | Kalshi | 592 | $6.5K | 30 | Buy days |
Trades and USD turnover summed over every contract of the series, from the recorded tick data; 30 days to 2026-11-01. Series without statistics in the window are omitted.
AI Models activity per venue
| Venue | Series | With stats | Trades (30d) | Turnover (30d) |
|---|---|---|---|---|
| Kalshi | 16 | 16 | 49,434 | $1.2M |
| predict.fun | 1 | 1 | 1,247 | $10.8K |
30 days to 2026-11-01, from our own daily statistics per contract.
AI Models — development since capture
126,461 trades and $3.1M of USD turnover across 17 series since 2026-08-09; busiest single series-day 2026-09-30 with $144.7K, biggest contract KXAISPIKE-27B-1540 (2026-10-28, $86.2K).
| Month | Series active | Days | Trades | Turnover (USD) | Avg / day | Contracts traded | L2 updates | vs previous month |
|---|---|---|---|---|---|---|---|---|
| Nov 2026 (month to date) | 16 | 1 | 1,137 | $20.3K | $20.3K | 63 | 278.4K | — |
| Oct 2026 | 17 | 31 | 54,953 | $1.4M | $46.6K | 2,156 | 11.8M | +16.9 % |
| Sep 2026 | 16 | 30 | 47,641 | $1.2M | $41.2K | 1,763 | 12.7M | +184.0 % |
| Aug 2026 | 10 | 23 | 22,730 | $434.7K | $18.9K | 1,159 | 6.8M | — |
AI Models series — trades and USD turnover summed per calendar month (UTC) over every contract, from our own daily statistics of the recorded tick data. Days = captured UTC days; a partial month shows no month-over-month change. Contracts traded and L2 (order-book) updates exist for days captured with the extended statistics.
AI Models — how the contracts resolved, last 8 days
Of 27 resolved AI Models contracts over the 8 captured days to 2026-10-30, 52 % closed Up and 48 % closed Down by their last print — 1,003 trades and $10.4K of turnover, about 7 trades per contract.
| UTC day | Contracts | Closed Up | Closed Down | Up share | Trades / contract | Turnover / contract | Turnover |
|---|---|---|---|---|---|---|---|
| 2026-10-30 | 21 | 1 | 0 | 100 % | 6 | $49.1 | $1.0K |
| 2026-10-29 | 17 | 2 | 0 | 100 % | 5 | $61.56 | $1.0K |
| 2026-10-28 | 22 | 1 | 2 | 33 % | 7 | $78.84 | $1.7K |
| 2026-10-27 | 22 | 2 | 2 | 50 % | 9 | $109.08 | $2.4K |
| 2026-10-26 | 13 | 0 | 0 | — | 3 | $97.3 | $1.3K |
| 2026-10-23 | 18 | 2 | 2 | 50 % | 3 | $23.81 | $428.51 |
| 2026-10-22 | 22 | 3 | 4 | 43 % | 12 | $91.12 | $2.0K |
| 2026-10-21 | 17 | 3 | 3 | 50 % | 4 | $26.49 | $450.28 |
| Total | 152 | 14 | 13 | 52 % | 7 | $68.16 | $10.4K |
Outcome implied by each contract's last recorded print (≥ 0.97 = Up, ≤ 0.03 = Down; the rest still open at the last print) — venues publish the official results, we record the market. Every contract of the day is counted for this series; trades and USD turnover come from the tick archive.
AI Models — turnover by weekday
Wednesdays are the busiest day at about $72.7K of USD turnover, averaged over the captured days of the last 30.
| Weekday (UTC) | Avg turnover / day | USD | Avg trades | Days |
|---|---|---|---|---|
| Monday | $44.0K | 1.9K | 4 | |
| Tuesday | $50.7K | 2.2K | 4 | |
| Wednesday | $72.7K | 2.0K | 4 | |
| Thursday | $42.4K | 1.7K | 4 | |
| Friday | $32.1K | 1.3K | 4 | |
| Saturday | $25.7K | 1.2K | 5 | |
| Sunday | $30.1K | 1.5K | 5 |
AI Models series — average USD turnover and trades per weekday over the captured days of the last 30, from our own daily statistics; weekdays without a captured day show —.
Rising AI Models series — 30 days vs the 30 before
| # | Series | Venue | Turnover (30d) | Prior 30d | Growth (avg / day) | Trades (30d) |
|---|---|---|---|---|---|---|
| 1 | Model hitting thresholds | Kalshi | $11.7K | $1.3K | +242.7 % | 816 |
| 2 | Top model | Kalshi | $25.9K | $20.0K | +29.5 % | 1,192 |
| 3 | Year-end top LLM | Kalshi | $1.0M | $1.0M | −0.7 % | 40,323 |
| 4 | Will any company achieve AGI before [date]? | Kalshi | $29.3K | $33.4K | −12.4 % | 698 |
| 5 | AI capability spike | Kalshi | $104.7K | $148.4K | −29.5 % | 718 |
Growth of average USD turnover per captured day, the 30 days to 2026-11-01 against the 30 days before; series need ≥ 10 captured days in both windows and ≥ $10K of turnover in the current one. From our own daily statistics.
AI Models prediction markets — trades and turnover per day
Every captured UTC day of the last 30, summed over the 17 AI Models series — how much these markets trade day by day, from our own tick capture.
| UTC day | Series with stats | Contracts traded | Trades | Turnover | Buy share | L2 updates |
|---|---|---|---|---|---|---|
| 2026-11-01 | 16 | 63 | 1,137 | $20.3K | 48 % | 278.4K |
| 2026-10-31 | 16 | 59 | 1,397 | $37.0K | 46 % | 260.6K |
| 2026-10-30 | 17 | 76 | 1,832 | $49.1K | 65 % | 512.6K |
| 2026-10-29 | 17 | 81 | 2,598 | $75.9K | 48 % | 474.0K |
| 2026-10-28 | 17 | 91 | 2,937 | $165.3K | 74 % | 373.4K |
| 2026-10-27 | 17 | 88 | 2,702 | $74.1K | 66 % | 529.2K |
| 2026-10-26 | 17 | 112 | 1,833 | $32.6K | 49 % | 270.0K |
| 2026-10-25 | 16 | 56 | 1,947 | $39.8K | 48 % | 276.6K |
| 2026-10-24 | 16 | 64 | 1,304 | $17.6K | 52 % | 221.1K |
| 2026-10-23 | 17 | 79 | 1,379 | $29.2K | 47 % | 245.9K |
| 2026-10-22 | 17 | 80 | 1,535 | $30.4K | 77 % | 423.6K |
| 2026-10-21 | 17 | 74 | 1,875 | $50.4K | 61 % | 434.5K |
| 2026-10-20 | 17 | 60 | 1,906 | $35.7K | 69 % | 474.4K |
| 2026-10-19 | 16 | 82 | 1,892 | $43.1K | 68 % | 450.4K |
| 2026-10-18 | 16 | 55 | 1,720 | $20.7K | 61 % | 471.8K |
| 2026-10-17 | 16 | 54 | 1,024 | $20.2K | 61 % | 326.5K |
| 2026-10-16 | 17 | 50 | 1,192 | $36.6K | 66 % | 407.9K |
| 2026-10-15 | 17 | 60 | 1,510 | $38.5K | 64 % | 306.8K |
| 2026-10-14 | 17 | 66 | 1,904 | $37.8K | 64 % | 564.0K |
| 2026-10-13 | 17 | 75 | 2,802 | $64.3K | 55 % | 535.8K |
| 2026-10-12 | 17 | 81 | 2,097 | $59.4K | 45 % | 388.0K |
| 2026-10-11 | 16 | 74 | 1,228 | $29.8K | 71 % | 218.4K |
| 2026-10-10 | 16 | 54 | 955 | $20.5K | 56 % | 159.6K |
| 2026-10-09 | 16 | 61 | 944 | $13.4K | 40 % | 218.7K |
| 2026-10-08 | 16 | 52 | 1,200 | $25.1K | 46 % | 338.4K |
| 2026-10-07 | 16 | 56 | 1,423 | $37.5K | 42 % | 372.9K |
| 2026-10-06 | 16 | 61 | 1,540 | $28.9K | 47 % | 392.2K |
| 2026-10-05 | 16 | 73 | 1,729 | $40.8K | 62 % | 351.9K |
| 2026-10-04 | 16 | 64 | 1,656 | $39.9K | 65 % | 251.0K |
| 2026-10-03 | 16 | 72 | 1,483 | $33.0K | 72 % | 329.5K |
Per UTC day, summed over every AI Models series that has statistics for that day, from every trade print of every contract in our tick archive. Trades and USD turnover are exact; contracts traded, buy share and L2 (order-book) updates exist for days captured with the extended statistics — earlier days show —.
Most traded AI Models contracts on 2026-11-01
The single event contracts with the most USD turnover on the newest captured day — with open and close price and the outcome implied by the last print.
| Contract | Series | Venue | Open → Close | Implied outcome | Trades | Turnover | |
|---|---|---|---|---|---|---|---|
| KXLLM1-26DEC31-A | Year-end top LLM | Kalshi | 0.391 → 0.396 | open | 289 | $9.4K | buy file → |
| KXLLM1-26DEC31-GOOG | Year-end top LLM | Kalshi | 0.527 → 0.481 | open | 336 | $4.9K | buy file → |
| KXLLM1-26NOV02-GOOG | Year-end top LLM | Kalshi | 0.990 → 0.990 | Yes | 23 | $2.2K | buy file → |
| KXLLM1-26DEC31-OAI | Year-end top LLM | Kalshi | 0.072 → 0.070 | open | 79 | $804.7 | buy file → |
| KXAGICO-COMP-27Q2 | Will any company achieve AGI before [date]? | Kalshi | 0.300 → 0.310 | open | 3 | $248 | buy file → |
| KXAISPIKE-27-1550 | AI capability spike | Kalshi | 0.570 → 0.580 | open | 20 | $243.05 | buy file → |
| KXLLM1-26DEC31-XAI | Year-end top LLM | Kalshi | 0.017 → 0.018 | No | 23 | $219.13 | buy file → |
| KXLLM1-26NOV30-GOOG | Year-end top LLM | Kalshi | 0.730 → 0.700 | open | 19 | $208.8 | buy file → |
| KXAGICO-COMP-28Q1 | Will any company achieve AGI before [date]? | Kalshi | 0.510 → 0.520 | open | 2 | $193.5 | buy file → |
| KXAISTREAMSERIES-27-NET | Will a major streaming platform offer a fully AI-generated scripted series by June 2026? | Kalshi | 0.080 → 0.200 | open | 14 | $178.46 | buy file → |
Single contracts ranked by USD turnover on 2026-11-01 (UTC), from every trade print in our tick archive. Prices quote 0..1; the implied outcome reads the last print of the day (≥ 0.97 Yes/Up, ≤ 0.03 No/Down) — venues publish the official result. Each contract’s day file is sold individually, the whole series day as a bundle.
Kalshi AI Models series
Largest archive first. Each series page lists its exact coverage, day count and recent days; day bundles are bought in the shop calendar.
| Series | Venue | Interval | Days | Coverage | Size | Price/day | |
|---|---|---|---|---|---|---|---|
| Year-end top LLM | Kalshi | event | 243 | March 2026 – present | 840 MB | €1 | Buy days |
| Will any company achieve AGI before [date]? | Kalshi | event | 243 | March 2026 – present | 178 MB | €1 | Buy days |
| Top model | Kalshi | monthly | 243 | March 2026 – present | 170 MB | €1 | Buy days |
| Claude Model Release | Kalshi | event | 227 | March 2026 – present | 141 MB | €1 | Buy days |
| AI Coding | Kalshi | event | 238 | March 2026 – present | 128 MB | €1 | Buy days |
| Humanity's Last Exam | Kalshi | event | 166 | May 2026 – present | 89.9 MB | €1 | Buy days |
| Best AI coding model | Kalshi | event | 243 | March 2026 – present | 78.9 MB | €1 | Buy days |
| Top-ranked AI | Kalshi | event | 243 | March 2026 – present | 66.5 MB | €1 | Buy days |
| AI capability spike | Kalshi | event | 243 | March 2026 – present | 62.6 MB | €1 | Buy days |
| Model hitting thresholds | Kalshi | event | 243 | March 2026 – present | 42.0 MB | €1 | Buy days |
| Will a major streaming platform offer a fully AI-generated scripted series by June 2026? | Kalshi | event | 243 | March 2026 – present | 42.0 MB | €1 | Buy days |
| What hardware will openAI/Jony Ive launch | Kalshi | event | 243 | March 2026 – present | 33.9 MB | €1 | Buy days |
| ANTHROPIC Sector in S&P | Kalshi | event | 152 | June 2026 – present | 23.7 MB | €1 | Buy days |
| Anthropic | Kalshi | event | 152 | June 2026 – present | 17.4 MB | €1 | Buy days |
| OpenAI | Kalshi | event | 152 | June 2026 – present | 11.0 MB | €1 | Buy days |
| OPENAI Sector in S&P | Kalshi | event | 152 | June 2026 – present | 9.75 MB | €1 | Buy days |
predict.fun AI Models series
Largest archive first. Each series page lists its exact coverage, day count and recent days; day bundles are bought in the shop calendar.
| Series | Venue | Interval | Days | Coverage | Size | Price/day | |
|---|---|---|---|---|---|---|---|
| ZHIPU Up/Down 15m | predict.fun | 15m | 14 | October 2026 – present | 16.4 MB | €1 | Buy days |
AI Models — market types across 16 series
1,237 contracts listed as of 2026-11-02, read from each contract's venue identifier — Kalshi lists each type as its own series, Polymarket, predict.fun and OG.com one family per league.
| Market type | Contracts listed (all time) | Listed last 30 days | Share |
|---|---|---|---|
| strike ladder | 17 | — | 1 % |
| moneyline / winner | 8 | — | 1 % |
| other | 1,212 | 121 | 98 % |
| Total | 1,237 | 121 | 100 % |
AI Models series— market type read from each contract's venue identifier (ticker suffix on Kalshi, slug suffix on Polymarket, market description on predict.fun and OG.com), counted over every contract listed in the family as of 2026-11-02; “listed last 30 days” counts contracts whose venue settlement time falls in that window. Every one of them is in the day bundle of its day.
Why buy AI Models prediction-market data here
Four things every page on this site is built on — and the reason the numbers above exist at all.
We record everything
The complete public feed of each venue as it was published: every Level-2 snapshot and update at the venue's full book depth, every trade with its aggressor side, every quote, funding, mark-price and liquidation event — for every instrument the venue lists, every UTC day since we added the venue. Nothing sampled, no top-N cut, no on-demand capture.
Institutional grade
Captured co-located at the venue with the exchange timestamp and our receive timestamp in integer nanoseconds, an event-id chain that makes any gap visible, and one normalized schema across 35+ venues — the same capture our own high-frequency trading engine and enterprise feeds run on.
€1 per instrument-day
Any instrument-day is €1, series-day bundles start at €1 — no subscription, no minimum order, no tiers to unlock. Credit packs lower the effective price and never expire, and every venue has free full-day samples to test against first.
Self-service for everyone
Pick the days in the Data Shop, pay by card as a guest and download immediately — no sales call, no enterprise contract, no KYC. Coding agents buy the same files through the MCP server, and the free Agent Skill teaches them the format.
Agents: /llms.txt · MCP server /mcp · every page as markdown via Accept: text/markdown
AI Models prediction-market data — FAQ
Does Kalshi have AI model markets?
Yes — year-end top LLM, top-ranked and top-coding-model series (Chatbot Arena based), Humanity’s Last Exam score ladders, AGI-by-date, model-release timing, capability-spike and hardware-launch markets. Polymarket’s "which company has the best AI model" families are archived under their own slugs.
How is AI prediction-market data sold?
One purchase covers every contract of the series for one UTC day — every trade print and the complete Level-2 book per contract, captured co-located with nanosecond timestamps — €1 per series-day, bought as a guest with instant download.
How far back does the AI Models data go?
Coverage runs March 2026 – present — 3,440 series-days across 17 series on Kalshi & predict.fun. Each series page lists its exact span.
Can I get AI Models prediction-market data as CSV or Parquet?
Yes — trades export as gzipped CSV or Parquet at no extra cost, with integer microsecond UTC timestamps that load straight into pandas, polars or DuckDB. Kalshi and predict.fun record full order-book depth without a separate BBO stream, so book data ships in the native tick file — the free reader script converts it locally. See the formats overview.
How much does AI Models prediction-market data cost?
€1 per series-day bundle — every contract of the series for that UTC day — no subscription, no minimum order. Pick days in the shop calendar, pay as a guest, download immediately.
Do I need an account or a sales call?
No. Pick the AI Models series-days in the Data Shop, pay by card as a guest and download immediately — no subscription, no minimum order, no KYC, no enterprise contract, no sales call. Coding agents buy the same bundles through the MCP server.
How is the AI Models data captured?
Co-located at the venue, message by message: every Level-2 update and every trade print carries the exchange timestamp and our receive timestamp in integer nanoseconds, an event-id chain makes any gap visible, and one normalized schema across 35+ venues means a parser written for Binance reads these files unchanged. It is the same capture our own high-frequency trading engine runs on — nothing sampled, no top-N cut.
Which market types are in the AI Models data?
As of 2026-11-02 the 16 AI Models series with a market-type snapshot had listed 1,237 contracts: strike ladder 17 · moneyline / winner 8 · other 1,212 — counted from every contract file of each family across Kalshi, predict.fun. Kalshi splits these into separate series; Polymarket, predict.fun and OG.com list one family per league. Each series page shows its own mix.
What was the busiest AI Models trading day?
2026-09-30 — $144.7K of USD turnover over 3,200 trades in a single series-day. Since our statistics began on 2026-08-09, 17 AI Models series recorded 126,461 trades and $3.1M in total; the biggest single contract was KXAISPIKE-27B-1540 on 2026-10-28 at $86.2K. Wednesdays are the busiest weekday at about $72.7K per day.
How often did AI Models contracts resolve Up vs Down recently?
52 % of 27 resolved contracts closed Up and 48 % closed Down by their last recorded print over the 8 captured days to 2026-10-30, about 7 trades per contract. Implied from the market (≥ 0.97 / ≤ 0.03), not the official settlement.