MiniMax 2.7: GLM-5 at 1/3 cost SOTA Open Model

报道来源 [4]

1. Smol AINews TIER_1 (CA) · 2026-03-18 05:44

   MiniMax 2.7: GLM-5 at 1/3 cost SOTA Open Model

   **MiniMax M2.7** is the headline model release, described as a "self-evolving agent" with strong performance metrics including **56.22% on SWE-Pro**, **57.0% on Terminal Bench 2**, and parity with **Sonnet 4.6**. It features recursive self-improvement in skills, memory, and archi…

2. Smol AINews TIER_1 English(EN) · 2026-02-11 05:44

   Z.ai GLM-5: New SOTA Open Weights LLM

   **Zhipu AI** launched **GLM-5**, an **Opus-class** model scaling from **355B to 744B parameters** with **DeepSeek Sparse Attention** integration for cost-efficient long-context serving. GLM-5 achieves **SOTA on BrowseComp** and leads on **Vending Bench 2**, focusing on office pro…

3. Smol AINews TIER_1 Nederlands(NL) · 2025-07-28 05:44

   GLM-4.5: Deeper, Headier, & better than Kimi/Qwen/DeepSeek (SOTA China LLM?)

   **Z.ai** (Zhipu AI) released the **GLM-4.5-355B-A32B** and **GLM-4.5-Air-106B-A12B** open weights models, claiming state-of-the-art performance competitive with **Claude 4 Opus**, **Grok 4**, and OpenAI's **o3**. These models emphasize token efficiency and efficient reinforcement…

4. dev.to — LLM tag TIER_1 English(EN) · RunC.AI Offical · 2026-05-09 07:47

   SGLang vs vLLM: Which LLM Serving Framework Should You Use?

   <p><em>Originally published at <a href="https://blog.runc.ai/sglang-vs-vllm/" rel="noopener noreferrer">https://blog.runc.ai/sglang-vs-vllm/</a>.</em></p> <h2 id="key-takeaways">Key Takeaways</h2> <ul> <li> <code>vLLM</code> is still the default starting point for many teams beca…