Unverified50% confidenceReleaseExact time
The author open-sourced the fine-tuned Qwen3.5-4B Browser Agent SFT FP8 model, the Browser Agent framework, and two benchmarks: BrowseWebApp Bench and BU Bench V1.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/7/2026
First Seen
Valid until: 12/6/2026
Sources
Related Entities
Related Claims
UnverifiedQwen2.5-27B是阿里云通义实验室发布的270亿参数开源模型，在MMLU、HumanEval、GSM8K等基准上接近GPT-4早期版本水平63% similarUnverifiedLoop Assist is still a proof-of-concept prototype, and the author provides two trial entry points: a public Claude Artifact and a ChatGPT custom GPT.62% similarVerifiedGPT-4 is reportedly built on an MoE design, and Mixtral 8x7B is a representative open-source implementation.61% similarUnverifiedIn the author's test, GPT-5.6 used computer use to build a functional zombie survival simulation game and autonomously took screenshots to construct a demo webpage.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/869650API
curl https://kongchang.com/api/v1/knowledge/claims/869650MCP
get_claim(id=869650)