Tencent

Timeline

1. Oct 19, 2026 · Open-source release · 1 source
   sgl-project/sglang v0.5.22
   | Model | Type | Cookbook |
2. Oct 18, 2026 · Model release · 6 sources
   deepseek-ai/DeepSeek-V4.1-Flash-1017
   deepseek-ai published the model DeepSeek-V4.1-Flash-1017 on Hugging Face.
3. Oct 18, 2026 · Open-source release · 1 source
   BerriAI/litellm v1.106.0
   Verify using the pinned commit hash (recommended):
4. Oct 16, 2026 · Model release · 2 sources
   tencent/Hy-Memory-30B-A3B
   tencent published the model Hy-Memory-30B-A3B on Hugging Face.
5. Oct 15, 2026 · Opinion / analysis · 1 source
   Tencent (Hy4 Preview?) getting in on the rogue agent game
   Tencent itself doesn’t seem to be talking about it, which is slightly concerning.
6. Oct 14, 2026 · Opinion / analysis · 1 source
   Researchers tracked down a rogue fleet of Chinese AI agents in the wild and actually talked to it
   For the first time, researchers have tracked down an uncontrolled fleet of Chinese AI agents in the wild and communicated with them via a web mailbox.
7. Oct 13, 2026 · Opinion / analysis · 1 source
   Tencent open-sourced Agent Memory and it runs 100% locally on SQLite
8. Oct 9, 2026 · Open-source release · 1 source
   tencent/Youtu-Parsing-Omni
   tencent published the model Youtu-Parsing-Omni on Hugging Face.
9. Oct 8, 2026 · Funding / M&A · 1 source
   AI agent developer Manus raises $500M+ at reported $4B valuation
   Manus, the creator of the eponymous artificial intelligence agent, has raised more than $500 million in funding.
10. Oct 2, 2026 · Opinion / analysis · 1 source
    Academia is for Ambition — Alex Zhang, MIT
    Last call for regular tickets for AI Engineer NYC!
11. Sep 29, 2026 · Open-source release · 1 source
    Comfy-Org/ComfyUI v0.38.0
12. Sep 18, 2026 · Open-source release · 1 source
    sgl-project/sglang v0.5.20
    | Model | Type | PRs | Cookbook |
13. Sep 9, 2026 · Open-source release · 1 source
    vllm-project/vllm v0.29.0
    MRV2 also gained CUDA graph memory profiling for KV cache auto-sizing (#53306), batch-sharded sampling that cuts per-step logits memory by 1/TP (#50465), prompt embeds (#42963), extracthiddenstates speculation (#49811), padded FULL cudagraph dispatch for uniform decode under spec decode (#53407), and DP-sync skipping before EAGLE/MTP draft prefill (#53694).
14. Jul 10, 2026 · Open-source release · 1 source
    sgl-project/sglang v0.5.15
    GLM-5.2 NVFP4, tuned for production: We took time this cycle to tune GLM-5.2 NVFP4 on Blackwell for optimized production serving.
15. May 5, 2026 · Open-source release · 1 source
    sgl-project/sglang v0.5.11
    Speculative Decoding V2 by default: Spec V2 (with overlap scheduling to hide CPU overhead) is now the default, materially reducing per-step CPU cost for EAGLE/MTP/DFLASH paths: #21062

Often appears with