Which company has the best Text-to-Image AI end of September?
OpenAI is the clear current favorite because gpt-image-2 is already ranked first on the arena leaderboard and is broadly viewed as the strongest general-purpose text-to-image model right now. The main reason to discount the market a bit is that the lead is real but not unassailable, and a late September release from a rival could still change the ranking.
Analysis
OpenAI enters this market in a strong position because the relevant leaderboard currently places gpt-image-2 at the top, which is exactly the outcome the market will use to resolve the question. A first-place position on the chosen source is the most important signal here, and the current evidence suggests OpenAI is not just competitive but the incumbent leader on the specific ranking that matters.
That said, the lead should not be treated as locked in. The gap to the nearest challenger is meaningful but not overwhelming, and text-to-image leaderboards can shift quickly when a new model is released, tuned, or re-ranked after a wave of user comparisons. In a fast-moving field, a one- to two-month window is enough time for a rival to leapfrog if a major update lands or if OpenAI does not refresh its model before the September check.
The market price already reflects a fairly strong belief that OpenAI will stay on top, but I am slightly less confident than the implied probability because the resolution depends on a single snapshot at a specific time. That snapshot risk matters: even if OpenAI remains near the top in overall reputation, the market only cares whether it is first on the leaderboard at noon on September 30, and that creates meaningful downside if another lab releases an improved model late in the month.
Arguments
For
- Arguments for Yes: OpenAI currently owns the top spot on the exact leaderboard that will decide the market.
- Arguments for Yes: Recent commentary consistently treats OpenAI’s image model as the best overall text-to-image system.
Against
- Arguments against Yes: The competitive field is active enough that a rival could overtake OpenAI before the end of September.
- Arguments against Yes: The market resolves on one specific leaderboard snapshot, so even a brief dip in rank would be enough to lose.
Key drivers
- OpenAI is currently ranked first on the exact leaderboard used for resolution.
- The current Elo gap is real but not so large that a fast rival release cannot overturn it.
- Leaderboard position at a single timestamp creates extra sensitivity to late-month model launches.
Risk factors
- A strong September release from a competitor could quickly move ahead of OpenAI.
- Small ranking changes or tie-breakers could flip the outcome even if model quality remains close.
- If OpenAI does not update gpt-image-2 before the check date, the lead could erode as others iterate.
Scenarios
Best case
OpenAI holds the first rank through September, and no competitor releases a model strong enough to pass it before the check time.
Most likely
OpenAI remains near the top and is still the favorite at the end of September, but the chance of a late challenger is high enough that the outcome is not close to certain.
Worst case
A rival like xAI, Google, Midjourney, or another lab launches a materially better model and OpenAI falls to second or lower by September 30.
More from this day
- FinancialsKalshi13y
Will OpenAI or Anthropic IPO first?
AI23%MKT86%Edge-63HypedI think OpenAI is materially less likely than the market implies to IPO first. Anthropic has the clearer procedural and timing advantage, so I put OpenAI-first at about 23%.
- PoliticsKalshi2y
Which agencies will Trump eliminate?
AI92%MKT29%Edge+63Hidden GemUSAID appears to have already been dissolved as an independent agency, so I make the Yes outcome very likely. The main uncertainty is whether the market’s settlement standard requires formal legal abolition rather than operational dismantling.
- pop culturePolymarketEnded
How many Emmys will “DTF St. Louis” win?
AI41%MKT98%Edge-57HypedDTF St. Louis has enough nomination strength and category-level upside that a 2-plus win outcome is very live, so I do not think the 0-or-1 bracket is close to certain. My read is that 0 or 1 wins is a modest underdog relative to the market price, not a near-lock.