Alibaba’s Qwen3.8-Max, a 2.4-trillion-parameter MoE model with 95 billion active parameters and a 1-million-token context window, launched on August 3, 2026, marking the first open-weights release at the Max tier and immediate availability via QwenCloud API. The model quickly gained ground on coding and long-horizon agent benchmarks, with a September 2 post-training snapshot (Qwen3.8-Max-0902) elevating it to first place on Code Arena through targeted reinforcement learning on enterprise workflows. This rapid iteration follows Qwen3.7 releases in May–June and trails Moonshot’s Kimi K3, reflecting Alibaba’s push to close gaps with Anthropic’s Fable 5 and OpenAI models while emphasizing inference efficiency and multimodal capabilities. Traders monitoring a 3.9+ successor should watch for further snapshots, Qwen4 architecture previews like the recent Flash-Next model, and competitive benchmark shifts that could accelerate or delay the next flagship.
Ringkasan eksperimental yang dihasilkan AI dengan referensi data Polymarket. Ini bukan saran trading dan tidak berperan dalam bagaimana pasar ini diselesaikan. · DiperbaruiSeptember 30
14%
October 31
60%
$8,214 Vol.
September 30
14%
October 31
60%
A qualifying model must have a name or model identifier that includes "Qwen" and "Max," be the flagship model of Alibaba's core Qwen Max tier, and be designated as version 3.9 or higher, regardless of capitalization or surrounding prefixes, suffixes, dates, or descriptors. For example, a version 3.9 or higher named in the same manner as Qwen3.7-Max or Qwen3.8-Max-Preview would qualify, including a new whole-number generation such as Qwen4-Max, while models whose name does not include "Max" or which retain a version designation below 3.9, such as Qwen3.8-Max, including its stable, open-weight, or snapshot re-releases, or Qwen3.8-Max-Preview, will not qualify. Derivative lines such as VL, Coder, Omni, or other specialized variants, such as qwen-vl-max, will not qualify.
A qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must either be clearly defined and publicly announced by Alibaba as accessible to the general public, or otherwise be made publicly accessible and explicitly labeled on the company's official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify.
The primary resolution source for this market will be official information from Alibaba, with additional verification from a consensus of credible reporting.
Pasar Dibuka: Aug 17, 2026, 9:42 AM ET
Resolver
0x65070BE91...A qualifying model must have a name or model identifier that includes "Qwen" and "Max," be the flagship model of Alibaba's core Qwen Max tier, and be designated as version 3.9 or higher, regardless of capitalization or surrounding prefixes, suffixes, dates, or descriptors. For example, a version 3.9 or higher named in the same manner as Qwen3.7-Max or Qwen3.8-Max-Preview would qualify, including a new whole-number generation such as Qwen4-Max, while models whose name does not include "Max" or which retain a version designation below 3.9, such as Qwen3.8-Max, including its stable, open-weight, or snapshot re-releases, or Qwen3.8-Max-Preview, will not qualify. Derivative lines such as VL, Coder, Omni, or other specialized variants, such as qwen-vl-max, will not qualify.
A qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must either be clearly defined and publicly announced by Alibaba as accessible to the general public, or otherwise be made publicly accessible and explicitly labeled on the company's official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify.
The primary resolution source for this market will be official information from Alibaba, with additional verification from a consensus of credible reporting.
Resolver
0x65070BE91...Alibaba’s Qwen3.8-Max, a 2.4-trillion-parameter MoE model with 95 billion active parameters and a 1-million-token context window, launched on August 3, 2026, marking the first open-weights release at the Max tier and immediate availability via QwenCloud API. The model quickly gained ground on coding and long-horizon agent benchmarks, with a September 2 post-training snapshot (Qwen3.8-Max-0902) elevating it to first place on Code Arena through targeted reinforcement learning on enterprise workflows. This rapid iteration follows Qwen3.7 releases in May–June and trails Moonshot’s Kimi K3, reflecting Alibaba’s push to close gaps with Anthropic’s Fable 5 and OpenAI models while emphasizing inference efficiency and multimodal capabilities. Traders monitoring a 3.9+ successor should watch for further snapshots, Qwen4 architecture previews like the recent Flash-Next model, and competitive benchmark shifts that could accelerate or delay the next flagship.
Ringkasan eksperimental yang dihasilkan AI dengan referensi data Polymarket. Ini bukan saran trading dan tidak berperan dalam bagaimana pasar ini diselesaikan. · Diperbarui

Hati-hati dengan link eksternal.
Hati-hati dengan link eksternal.
Pertanyaan yang Sering Diajukan