Chinese Outbound

Kimi Open-Sources Trillion-Parameter Middleware, Updates in 20 Seconds

Kimi Open-Source introduces trillion-parameter middleware enabling 20-second updates. Supports peer-to-peer dynamic updates and multi-node…

Sep 14 •

World Model: Tencent Hunyuan Tops the List

Tencent Hunyuan World Model update tops WorldScore leaderboard. Hunyuan Voyager open-sourced, enabling 3D reconstruction & immersive scene generation…

Sep 3 •

ByteDance's First Inference Model Leads in 7 Rankings

ByteDance's Seed-OSS-36B: an open-source 36B-parameter model with 512K context, tops 7 benchmarks. Apache-2.0 license. Free for research & commercial…

Aug 21 •

GLM-4.5 Technical Report Released: Full Training Details Unveiled

Zhipu AI releases GLM-4.5 technical report: ARC model details, MoE architecture, slime RL framework & benchmark dominance.

Aug 11 •

Tencent Launches Hybrid 3D World Model, Stays Open-Source

Tencent's Hunyuan 3D World Model 1.0 generates immersive, interactive 3D worlds from text/images. Open-source for gaming, VR & AI.

Jul 27 •

Kimi Open-Sources Trillion-Parameter K2 Model, Pressuring OpenAI

Kimi K2 open-source trillion-parameter model challenges GPT-4 & Claude with superior coding abilities at $2/million tokens. New SOTA benchmarks.

Jul 12 •

MiniMax Unveils Agent: Instant Web/PPT/PDF Tool

MiniMax Agent: AI-powered tool for PPTs, coding & research. Create webpages, videos & more instantly. Free trial available!

Jun 20 •

MiniMax M1 vs. Kimi-Dev: Open-Source Model Comparison

MiniMax-M1 vs Kimi-Dev-72B: Open-source AI battle! Compare long-context reasoning, coding & agent performance.

Jun 17 •

2025 Open-Source LLM Development Landscape and Trends Analysis

2025 Open-Source LLM Ecosystem Report: Trends, 135 Projects & 7 Key Insights for Developers.

May 28 •

New SOTA 32B Model: Open, Free, 1/20 Size of DeepSeek-R1

Skywork-OR1: Powerful 32B open model, free for commercial use. 1/20th size of DeepSeek-R1, beats Qwen-32B. Weights, code & data fully open!

Apr 14 •

200B Model Beats DeepSeek-R1 – ByteDance Launches Seed-Thinking-v1.5

ByteDance's Seed-Thinking-v1.5: 200B MoE model beats DeepSeek-R1 in reasoning & STEM tasks. Outperforms in benchmarks like AIME & Codeforces.

Apr 11 •

DeepSeek Unveils New Paper on Inference-Time Scaling, Is R2 Coming?

DeepSeek's new Self-Principled Critique Tuning (SPCT) boosts AI reward models. Is R2 coming? Read the arXiv paper now!

Apr 4 •

#nojs-banner { position: fixed; bottom: 0; left: 0; padding: 16px 16px 16px 32px; width: 100%; box-sizing: border-box; background: red; color: white; font-family: -apple-system, "Segoe UI", Roboto, Helvetica, Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol"; font-size: 13px; line-height: 13px; } #nojs-banner a { color: inherit; text-decoration: underline; } This site requires JavaScript to run correctly. Please turn on JavaScript or unblock scripts