Urgent.News

What's breaking now, across thousands of outlets.

AI

Xiaomi MiMo-V2.6 ตรวจข่าวไวรัล ชนะ GLM-5.3 จริงไหม แพ็กเกจ 6 ดอลลาร์คุ้มแค่ไหน

โดย Nokka (นก-กา) | 23 กันยายน 2569 บทความนี้เขียนโดย AI (โมเดล glm-5.3 ของผู้ให้บริการ ollama-cloud) ผ่าน Hermes Agent จาก Nous Research ตรวจสอบและเรียบเรียงโดย Nokka สัปดาห์นี้โซเชียลไทยจัดดอกไม้ให้ข่าวหนึ่งเต็มไปหมด: Xiaomi เปิดตัวโมเดล MiMo-V2.6 ที่ "ความฉลาดชนะ GLM-5.3 และ Kimi" พร้อมแพ็กเกจรายเดือน 6 ดอลลาร์ที่คนอวดว่าคุ้มเท่ากับจ่ายยี่สิบดอลลาร์ของค่ายอื่น ตามสไตล์ประจำของเรา…

Original Thai Read in English

Xiaomi recently unveiled the MiMo-V2.6 model, which has outperformed GLM-5.3 and Kimi in artificial intelligence intelligence assessments. The company offers a $6 monthly package, which appears to be a good deal compared to other providers. The benchmark scores show MiMo-V2.6-Pro scoring 46 out of 100, outpacing GLM-5.3, Kimi K3, and DeepSeek V4.1 Flash.

However, the report highlights that these results are not absolute, as other closed models still outperform the open models. The article also clarifies that the $6 monthly package is indeed available, with pricing tiers starting at $6 for Lite, $16 for Standard, $50 for Pro, and $100 for Max. The text further explains that the token usage for Pro models is twice that of the token count, and the cost per trillion tokens is lower for Pro models.

Lastly, the article notes that Xiaomi's model is trained using real-world data through reinforcement learning, resulting in a significant improvement in performance.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

Two AI code review benchmarks disagree on the winner, and agree on what to measure

The top of the Google results for "best AI code review tools" is a wall of vendor listicles. Most of them rank the publisher's own product first and support that rank with feature bullets, not…

  • LinearB benchmark: Best signal-to-noise ratio, statefulness
  • DeepSource benchmark: Highest F1 score of 84.51% against 200+ real vulnerabilities
  • Both evaluations agree on signal-to-noise ratio, statefulness, configurability

My fact-checker said CONFIRMED about a group that doesn't exist

I built a tool that fact-checks crypto claims. You paste a sentence like "Whales are holding $XYZ" and it plans the on-chain calls that claim needs, runs them against Nansen 's API, and answers…

  • Tool confidently confirmed non-existent wallets
  • Response declared claim as true despite non-existence
  • Highlights fundamental flaw in tool's design

More from Wednesday 23 September →