# Which AI company will have the best coding model on Dec 31, 2026

> Anthropic leads at 60%, runner-up 30% across 4 winner-take-all outcomes — refreshed 30 min ago.

URL: https://simplefunctions.dev/odds/codingmodel
Updated: 2026-06-08T07:20:08.898Z
Category: technology
Status: active
Closes: 2026-12-31

## Headline

- Leader: Anthropic at 60%
- Runner-up: OpenAI at 30%
- Outcomes: 4 (winner-take-all)
- Venue: Kalshi (4 contracts)
- 24h volume: $5K

## Bound contracts (4)

| Outcome | Price | 24h | Volume | Venue | Slug |
|---|---|---|---|---|---|
| Anthropic | 60¢ | +2pp | $1K | kalshi | /markets/which-ai-company-will-have-the-best-coding-model-o-kalshi-kxcodingmodel-26dec-anth |
| OpenAI | 30¢ | +1pp | $2K | kalshi | /markets/which-ai-company-will-have-the-best-coding-model-o-kalshi-kxcodingmodel-26dec-open |
| xAI | 5¢ | −1pp | $1K | kalshi | /markets/which-ai-company-will-have-the-best-coding-model-o-kalshi-kxcodingmodel-26dec-xai |
| Google | 5¢ | −1pp | $451 | kalshi | /markets/which-ai-company-will-have-the-best-coding-model-o-kalshi-kxcodingmodel-26dec-goog |

## 30-day trajectory

| Day | Anthropic | OpenAI | xAI |
|---|---|---|---|
| 2026-05-09 | 57 | 18 | 6 |
| 2026-05-24 | 58 | — | 3 |
| 2026-05-25 | 59 | 21 | — |
| 2026-05-31 | 54 | 28 | 7 |
| 2026-06-01 | 58 | 30 | — |
| 2026-06-07 | 62 | 28 | 5 |
| 2026-06-08 | — | 29 | — |

_31 days of price history captured. Each row is the daily mean of intraday 5-min captures._

## What moved the line

- 2026-06-01 · Anthropic +4pp 54→58¢ · kalshi
- 2026-06-05 · Anthropic +4pp 55→59¢ · kalshi
- 2026-06-02 · xAI −3pp 7→4¢ · kalshi

## Analysis

This probability represents the market's assessment that a single AI company will have demonstrably the best coding model by year-end 2026. The 28-percentage-point gap between Polymarket (49%) and Kalshi (21%) reflects disagreement on how to evaluate and compare coding capabilities across different vendors. Current market pricing suggests OpenAI and Anthropic are primary contenders, though recent April assessments show fragmented views on which company leads in specialized AI tasks. The probability will likely shift based on new model releases, benchmark results from independent evaluators like LMSYS or academic leaderboards, and how "best" gets operationalized—whether measured by coding accuracy, speed, real-world adoption, or specific benchmarks. Key resolution catalysts include major model announcements (OpenAI, Anthropic, Google, Meta typically release updates in clusters), publication of standardized coding benchmarks, and community consensus on evaluation methodology.

### Key factors

- OpenAI's current coding model is priced at 3¢ on Polymarket for best-coding-model-end-of-April, suggesting market skepticism of its near-term position despite historically strong coding performance
- Anthropic shows strong positioning in math reasoning (81¢) but unclear coding-specific strength, creating uncertainty about cross-domain model capabilities
- Polymarket prices are 2-3x higher than Kalshi equivalents across similar contracts, indicating structural differences in how venues' participants assess model competition
- No single standardized benchmark is referenced in current contracts; resolution depends on how the prediction market operator defines and measures 'best' coding model
- Eight months remain for multiple companies to release new models, making current April 2026 assessments potentially outdated by year-end evaluation date

## Methodology

Headline is the **leader's price**, not an arithmetic mean — averaging disjoint winner-take-all outcomes is meaningless. Per-outcome prices come from the venue's last-traded mid; cross-venue values are simple means across contracts on each venue.

## How to use this data

- HTML: https://simplefunctions.dev/odds/codingmodel
- JSON: https://simplefunctions.dev/api/public/odds?slug=codingmodel

## License

CC-BY-4.0. Attribute "SimpleFunctions" with a link to https://simplefunctions.dev. See https://simplefunctions.dev/legal for terms.