Asia AI launches test US export curbs as Polymarket keeps Anthropic at 85.5%
Alvin Lang Jun 27, 2026 12:26
Two weeks after a U.S. order blocked global access to Anthropic’s Mythos and Fable models, China’s 360 unveiled the Tulongfeng tool and Sakana AI launched Fugu, pitching export-control resilience.
Asia’s New AI Model Launches Test Anthropic—Polymarket Still Prices Anthropic as the End‑July 2026 Favorite at 85.5%
New AI model launches in Asia are testing the dominance of top U.S. frontier systems just as U.S. export controls restrict access to Anthropic’s most capable models. On Polymarket’s "Which company has best AI model end of July?" contract, traders still price Anthropic as the clear favorite, even after a modest uptick in its implied odds.
Key Takeaways
- Polymarket prices Anthropic as the leader at 85.5% to have the best AI model by end-July 2026.
- The odds shift comes as Asian firms tout new models positioned against Anthropic-branded benchmarks amid U.S. export restrictions on Anthropic systems.
- The market resolves on 2026-07-31, with Anthropic up 1.5 percentage points from 84.0% in the latest snapshot.
A Chinese cybersecurity company, 360, reportedly unveiled an AI tool called Tulongfeng that it said could compete with Anthropic’s Mythos, a cybersecurity-focused model described as powerful enough that the Trump Administration has banned it and a more restricted version, Fable 5, from non-Americans. Earlier in the week, Tokyo-based startup Sakana AI launched a model called Fugu and said it can match leading systems such as Anthropic’s Fable 5 and Mythos Preview. Sakana said Fugu is built for agentic use cases, including orchestrating access to other models through their APIs, and framed it as a way to preserve frontier capability as export controls tighten. The launches come two weeks after a U.S. government order that blocked global access to Anthropic’s Mythos and Fable models, and Sakana said the timing was coincidental while marketing its product as reducing exposure to export controls. 360 also reportedly introduced another security tool, Yitianzhen, aimed at automating cyber defense and incident response, as its founder described vulnerability-finding AI as a strategic asset and warned of “one-way transparency” risks.
Polymarket Odds and Liquidity: $2.22M Traded as Anthropic Holds 85.5% vs Google 10.85% and OpenAI 3.6%
On Polymarket, the multi-outcome contract shows Anthropic at 85.5% Yes / 14.5% No, with about $2,223,710 traded, underscoring a heavily one-sided consensus around that outcome. Google is a distant second at 10.85% Yes / 89.15% No, while OpenAI is priced at 3.6% Yes / 96.4% No. The long tail is near-zero: xAI is 0.25% Yes / 99.75% No, and several others (including Alibaba, Meta, Microsoft, Baidu, and DeepSeek) sit around 0.15% Yes / 99.85% No. The tight clustering at the bottom suggests the market is treating the race as effectively between Anthropic and a small number of challengers heading into the July 31, 2026 resolution date.
Watch whether the leader’s implied odds hold above the mid-80s as liquidity rotates between Anthropic and Google, and whether the spread between the top two outcomes narrows ahead of the 2026-07-31 resolution.
Beyond the AI Race: Other High‑Volume Polymarket Contracts Traders Are Watching Across Geopolitics and Macro
Beyond the model leaderboard, traders are also spreading risk across adjacent tech timing bets and broader macro-geopolitical arcs on Polymarket. One of the most active is “GPT-5.6 released by...?”, where the leading outcome “July 31” is priced at 88.9% on $1,751,794 in volume after a 6.2-point move, highlighting how quickly sentiment can swing even in ostensibly date-driven contracts.
Odds Trend
| Window | Change (pp) |
|---|---|
| 24h | -2.5 |
| 7d | -2.5 |
By the Numbers
- Platform: Polymarket
- Market: Which company has best AI model end of July?
- Contract type: Price strike ladder: each rung has separate Yes/No; Yes means the spot price is above that USD strike at settlement.
- Resolution window: Jul 31, 2026 (UTC)
- Status: Active (open for trading)
- Volume: ~$2,223,710
Top strike rungs
| Strike | Yes | No |
|---|---|---|
| Anthropic | 85.5% | 14.5% |
| 10.8% | 89.2% | |
| OpenAI | 3.6% | 96.4% |
| xAI | 0.2% | 99.8% |
+11 more strikes not shown
Related Markets
- GPT-5.6 released by...? — July 31 89%