model-leak

Alibaba's QwenLM/qwen-code repository reportedly leaked references to upcoming models qwen4-max, qwen4-coder-plus, and qwen4-omni-120b ahead of any official announcement.

GPT-6.2 Astra appears in some Codex users' usage analytics, suggesting OpenAI may be preparing a staged rollout or A/B test of its next frontier model ahead of a possible launch.

A Hugging Face repo called burner22/numbat claims to contain an Anthropic internal training checkpoint, reportedly 14T parameters, 114 layers, and 1,224 MoE experts — suspected to be one of the largest leaks of closed-source model weights ever.

疑似 Muse Wonder 1.0 被曝光或发现，可能是一个新 AI 模型的首次现身。

Google's unreleased Gemini 4 Argon reportedly leaked and outperforms Fable 5.1, Opus 5.5 and GPT-6 Astra on 12 of 18 benchmarks, including a large lead in autonomous legal work (19.6% vs 5.4%).

Stepfun's new AI model, Step 5 Preview, has been leaked, showing a significant version jump from 3.7 to 5.0 with major performance improvements.

A reliable leaker has shared one-shot output samples from the Astra model operating at maximum effort, condensed into a video.

本文讨论了谷歌Gemini 3.8 Flash模型的泄露消息，指出该模型已在内部部署并进行了测试，性能可能接近Fable 5水平。

A potential new GPT-Image model, reportedly named "Mona-lisa-1", has appeared on the LMArena, with OpenAI SynthID watermarks detected in its outputs.