Genspark 与 Baseten 联合后训练开源模型，深度研究能力超过 Opus 5.5
Genspark 和 Baseten 把 Qwen3.5-397B 后训练成深度研究专用模型，内评测分 0.759，直接超过 Opus 5.5 的 0.738，现在已经跑在自家 Deep Research 产品里。做垂直场景落地的团队可以看看这条路：开源基座加定向后训练，单任务能打赢闭源旗舰。
Genspark 与 Baseten Applied Research 合作，基于开源权重模型 Qwen3.5-397B 做后训练，产出专攻深度研究任务的模型。在 Genspark 内部评测中，该模型得分为 0.759，超过 Opus 5.5 的 0.738，明显高于基座模型 Qwen3.5-397B 的 0.612。
技术上，团队没有从零训练，而是对 Qwen3.5-397B 进行后训练，把能力集中在 deep research 这一类多步检索与综合任务上。该模型现已上线，成为 Genspark Deep Research 智能体在生产环境中的默认模型。
这件事对开发者的参考意义在于：一个 397B 的开源基座模型，经过针对具体工作负载的后训练，可以在单一任务上压过更大的前沿闭源模型。用内部评测集驱动后训练、再直接替换生产默认模型，是一条成本可控的路径。
Another model from Genspark, partnership with Baseten! Beats frontior models at deep research! Baseten @baseten Genspark and Baseten Applied Research post-trained an open-weight model that beats Opus 5.5 at deep research. Starting from Qwen3.5-397B, our post-trained model scores 0.759 on @genspark_ai 's internal evals, ahead of Opus 5.5 (0.738) and well above the base model (0.612). It’s now the default model in production behind Genspark’s Deep Research agent. 🔗 View Quoted Tweet📊 1 ⚡
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Genspark, Baseten, Qwen3.5, 开源模型.
Source attribution: curated from Eric Jing (@ericjing_ai). Cross-source corroboration: 3 independent sources have covered this development (including Genspark, arXiv cs.LG), confirming industry-wide signal strength.
