AI News

Trending story

LG AI Research Institute opens the K-EXAONE 2.5 model API for free

5 reports1 discussions6 story updatesUpdated Nov 9 15:31

Understand this story

AI overview

English text is being prepared.

AI-generated from public reports

Latest development · Report

NVIDIA Holds NVIDIA AI Day in Seoul; South Korean Government, Samsung, LG Electronics, Hyundai Motor and Others Announce Cooperation

Developments

1 confirmed developments

  1. LG AI Research opens K-EXAONE 2.5 model API for free

    Source:Unite.AI

Reports and discussion timeline

Browse reports and discussion by time. Their evidence levels remain distinct.

Nov 9

  1. IT Home · Artificial Intelligence

    NVIDIA Holds NVIDIA AI Day in Seoul; South Korean Government, Samsung, LG Electronics, Hyundai Motor and Others Announce Cooperation

    On November 9, NVIDIA held NVIDIA AI Day in Seoul, South Korea, announcing cooperation with the South Korean government and multiple companies. NVIDIA expanded sovereign AI ecosystem cooperation with South Korea's Ministry of Science and ICT (MSIT), and is cooperating with four foundation model builders—SK Telecom, LG AI Research, Upstage, and Motif Technologies—to optimize inference using NVIDIA Dynamo. Samsung Electronics expanded cooperation in accelerated computing and AI physical technology, LG Electronics demonstrated the CLOiD robot running NVIDIA Isaac GR00T, and Kakao Mobility is advancing L4-level autonomous driving.

  2. Sina Tech

    NVIDIA Holds NVIDIA AI Day in Seoul; South Korean Government, Samsung, LG, Hyundai and Others Announce Multiple Collaborations

    On November 9, NVIDIA held NVIDIA AI Day in Seoul, expanding sovereign AI ecosystem cooperation with South Korea's Ministry of Science and ICT (MSIT), and building NVLink Fusion-based infrastructure together with SK Telecom and Rebellions. Samsung Electronics expanded its semiconductor cooperation with NVIDIA in accelerated computing and AI physical technologies, LG Electronics demonstrated a CLOiD robot manufacturing task driven by Isaac GR00T, and Hyundai Motor Group is exploring the NVIDIA NemoClaw blueprint and OpenShell runtime. LG AI Research also opened API access to K-EXAONE-2.5.

  3. NVIDIA:NewsroomFirst-party

    South Korea's AI ecosystem brings in NVIDIA OpenShell and Nemotron to strengthen agent security

    Hyundai Motor Group is exploring the NVIDIA NemoClaw blueprint and the NVIDIA OpenShell runtime, while South Korean AI security company AIM Intelligence has integrated its Starfort agent control platform with OpenShell, combining runtime access control with malicious instruction and data leakage detection capabilities. OpenShell is a core component of the NVIDIA Open Agent Safety Platform, providing sandboxed execution and policy-based controls. LG AI Research has also opened API access to K-EXAONE-2.5, developed based on NVIDIA Nemotron datasets (including Nemotron-Personas-Korea, which contains 7 million synthetic Korean user personas).

  4. English text is being prepared.

    LG AI Research Opens K-EXAONE 2.5 Model API for Free

    LG AI Research has announced that it is opening the K-EXAONE 2.5 model API to the public for free; developers, enterprises, and research institutions can all use it at no cost for service development and research. The model was further trained on top of the original version and possesses agentic capabilities; its performance on AutomationBench, SWE-Bench Pro, BrowseComp, and Claw-Eval is on par with DeepSeek-V4.1-Flash and GLM-5.3.

  5. IT Home · AIRelated discussion

    IT Home · AI joined this discussion

    This post contributes to heat; it is not a formal report or confirmed fact.

  6. Unite.AI

    LG AI Research Opens Free API for K-EXAONE 2.5 Model

    On November 9, 2026, LG AI Research announced that it is offering a free hosted API for K-EXAONE 2.5 through the EXAONE API Platform; developers, startups, enterprises, and research institutions can all apply for an API key for service development and research. The model is text-only and supports Korean and English. The documentation lists 31B activated parameters and a January 2026 knowledge cutoff. The interface is the OpenAI-compatible POST /v1/chat/completions, the model identifier is k-exaone_v2.5, and the combined input and max_tokens must not exceed 524,288 tokens. LG says that internal measurements on open datasets show it outperforms the previous-generation K-EXAONE 2.0 on all four agent benchmarks: SWE-Bench Pro, BrowseComp, AutomationBench, and Claw-Eval, and it also provides comparison data with DeepSeek-V4.1-Flash and GLM-5.3.

Heat trend for this story

Current
5
Period peak
19 Nov 9 13:00
24-hour change
-50%

Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.

Related stories

  1. Related storiesOpenAI launches GPT-6.1 Terra on ChatGPT Work and Codex
  2. Related storiesMeta launches Muse Wonder and Muse Wonder Full models
  3. Related storiesMistral releases Large 4 weights on Hugging Face
  4. Related storiesAnthropic Publicly Submits S-1 for Proposed Nasdaq IPO