{"id":6190,"date":"2026-09-28T12:05:25","date_gmt":"2026-09-28T03:05:25","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/publisher-china-telecom-en\/"},"modified":"2026-09-28T12:05:25","modified_gmt":"2026-09-28T03:05:25","slug":"publisher-china-telecom-en","status":"publish","type":"page","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-china-telecom-en\/","title":{"rendered":"China Telecom (Xing, TeleChat) Open Models: Model List, Licenses and Articles"},"content":{"rendered":"<h2>About China Telecom (Xing, TeleChat)<\/h2>\n<p><strong>China Telecom<\/strong> is a major Chinese telecommunications carrier. Its AI units have released the <strong>TeleChat<\/strong> language models on Hugging Face. China Telecom&#8217;s Institute of Artificial Intelligence (TeleAI) published TeleChat on the <a href=\"https:\/\/huggingface.co\/Tele-AI\">Tele-AI organization<\/a> (22), and since September 2026 its successor <strong>Xing<\/strong> has been released by China Telecom Artificial Intelligence Technology Co., Ltd. on the <a href=\"https:\/\/huggingface.co\/XingChen-AGI\">XingChen-AGI organization<\/a> (4).<\/p>\n<p>Its defining trait is <strong>training only on domestic Chinese compute<\/strong>. The TeleChat3 model card states that the series was trained entirely on domestic computing power, and the Xing4.0-29B-A4B model card calls it the first model of its scale trained entirely on Huawei&#8217;s Ascend NPUs with the MindSpore framework. What sets it apart from other publishers is producing results comparable to leading open models of the same class without NVIDIA GPUs.<\/p>\n<h2>What It Releases<\/h2>\n<ul>\n<li><strong>Xing4.0-29B-A4B<\/strong> (September 2026): an MoE with 29B total and 4B active parameters, with a 256K-token context (extensible to 512K). It is built for coding agents, and in the model card&#8217;s comparison it beats Qwen3.6-35B-A3B and Gemma4-26B-A4B on the terminal task benchmark Terminal-Bench 2.1 (<a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/28\/xing4-0-29b-a4b-china-telecom\/\">our article<\/a>). Official FP8 and GGUF builds were released at the same time.<\/li>\n<li><strong>TeleOCR<\/strong> (August 2026): a vision-language model for reading text from images, released from the XingChen-AGI organization.<\/li>\n<li><strong>TeleChat3<\/strong> (December 2025): an MoE with 105B total and 4.7B active parameters (105B-A4.7B-Thinking) and a 36B dense model (36B-Thinking). A coding-focused TeleChat3-Coder-36B-Thinking followed in January 2026.<\/li>\n<li><strong>TeleChat2 \/ TeleChat2.5 \/ T1<\/strong> (2024\u20132025): dense models from 3B to 115B. T1 is the reasoning-focused variant.<\/li>\n<li>Earlier releases include TeleChat (7B, 12B, 52B), the speech recognition model TeleSpeech-ASR and the image style transfer model TeleStyle.<\/li>\n<\/ul>\n<h2>Licenses<\/h2>\n<p>Xing4.0-29B-A4B, TeleOCR and TeleChat2-115B are under <strong>Apache-2.0<\/strong>, broadly usable including commercially. The TeleChat3 model card, by contrast, gives no license identifier and carries a statement asking users not to use the models for activities that harm national or social security or are illegal, and not to use them in internet services that have not passed security review and registration. Read each model&#8217;s license text before use. This summary is not legal advice.<\/p>\n<h2>Running Them Locally<\/h2>\n<ul>\n<li><strong>Xing4.0-29B-A4B is the first candidate for local use.<\/strong> The publisher provides its own GGUF build, and with 4B active parameters generation is light once the weights fit in memory. All 29B of the weights still have to fit in memory.<\/li>\n<li><strong>TeleChat3-36B is a dense model<\/strong>, so it generates more slowly than an MoE of similar size. The 105B-A4.7B model has a large total size, making memory tight on consumer hardware even when quantized.<\/li>\n<li><strong>Many Transformers examples use <code>trust_remote_code=True<\/code>.<\/strong> The TeleChat3 repositories include their own model definition code.<\/li>\n<\/ul>\n<p><em>Sources: <a href=\"https:\/\/huggingface.co\/XingChen-AGI\">the XingChen-AGI organization card<\/a>; the model cards of <a href=\"https:\/\/huggingface.co\/XingChen-AGI\/Xing4.0-29B-A4B\">XingChen-AGI\/Xing4.0-29B-A4B<\/a> and <a href=\"https:\/\/huggingface.co\/Tele-AI\/TeleChat3-36B-Thinking\">Tele-AI\/TeleChat3-36B-Thinking<\/a>; the model lists of the XingChen-AGI and Tele-AI organizations on Hugging Face (all as of 2026-09-28).<\/em><\/p>\n<h2>Our Coverage and Data<\/h2>\n<p>Local Model Watch has published 1 article(s) on China Telecom (Xing, TeleChat)&#8217;s own models, 0 on third-party fine-tunes and quantized builds based on them, and 0 from its official blog. The model list below is collected by code from Hugging Face every day. Part of our <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publishers-en\/\">publisher index<\/a>; model families link to their <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/models-en\/\">family pages<\/a>.<\/p>\n<h2>At a Glance<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Item<\/th>\n<th>Value<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Hugging Face organization<\/td>\n<td><a href=\"https:\/\/huggingface.co\/XingChen-AGI\">XingChen-AGI<\/a>, <a href=\"https:\/\/huggingface.co\/Tele-AI\">Tele-AI<\/a><\/td>\n<\/tr>\n<tr>\n<td>Models on Hugging Face<\/td>\n<td>26<\/td>\n<\/tr>\n<tr>\n<td>Articles on its own models<\/td>\n<td>1<\/td>\n<\/tr>\n<tr>\n<td>Articles on derived models<\/td>\n<td>0<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h2>Recent Models on Hugging Face<\/h2>\n<p>The publisher lists <strong>26 models<\/strong> in total on Hugging Face. Below are up to 15 of the most recent generative models (text, image, video, audio and so on), the scope of this site, per organization. Classifiers, feature extractors, research components and repositories without a model card are omitted, and the publisher&#8217;s own quantized builds and helper variants are folded into the row of the original model. Parameter counts are left blank for repositories that pack weights into integer types, where Hugging Face&#8217;s count is not the real parameter count.<\/p>\n<h3><a href=\"https:\/\/huggingface.co\/XingChen-AGI\">XingChen-AGI<\/a> (4 models)<\/h3>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Released<\/th>\n<th>Model<\/th>\n<th>Task<\/th>\n<th>Parameters<\/th>\n<th>License<\/th>\n<th>Official variants<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>2026-09-16<\/td>\n<td><a href=\"https:\/\/huggingface.co\/XingChen-AGI\/Xing4.0-29B-A4B\">XingChen-AGI\/Xing4.0-29B-A4B<\/a><\/td>\n<td>text generation<\/td>\n<td>31.2B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>GGUF, FP8<\/td>\n<\/tr>\n<tr>\n<td>2026-08-14<\/td>\n<td><a href=\"https:\/\/huggingface.co\/XingChen-AGI\/TeleOCR\">XingChen-AGI\/TeleOCR<\/a><\/td>\n<td>vision-language<\/td>\n<td>1.4B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h3><a href=\"https:\/\/huggingface.co\/Tele-AI\">Tele-AI<\/a> (22 models)<\/h3>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Released<\/th>\n<th>Model<\/th>\n<th>Task<\/th>\n<th>Parameters<\/th>\n<th>License<\/th>\n<th>Official variants<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>2026-06-10<\/td>\n<td><a href=\"https:\/\/huggingface.co\/Tele-AI\/TeleStyleV2\">Tele-AI\/TeleStyleV2<\/a><\/td>\n<td>vision-language<\/td>\n<td>\u2014<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2026-01-27<\/td>\n<td><a href=\"https:\/\/huggingface.co\/Tele-AI\/TeleStyle\">Tele-AI\/TeleStyle<\/a><\/td>\n<td>vision-language<\/td>\n<td>\u2014<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2025-10-21<\/td>\n<td><a href=\"https:\/\/huggingface.co\/Tele-AI\/MMPL\">Tele-AI\/MMPL<\/a><\/td>\n<td>image generation<\/td>\n<td>\u2014<\/td>\n<td>\u2014<\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2024-09-20<\/td>\n<td><a href=\"https:\/\/huggingface.co\/Tele-AI\/TeleChat2-115B\">Tele-AI\/TeleChat2-115B<\/a><\/td>\n<td>\u2014<\/td>\n<td>113.3B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2024-06-28<\/td>\n<td><a href=\"https:\/\/huggingface.co\/Tele-AI\/TeleChat-1B\">Tele-AI\/TeleChat-1B<\/a><\/td>\n<td>text generation<\/td>\n<td>\u2014<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2024-03-19<\/td>\n<td><a href=\"https:\/\/huggingface.co\/Tele-AI\/TeleChat-12B\">Tele-AI\/TeleChat-12B<\/a><\/td>\n<td>text generation<\/td>\n<td>\u2014<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>int4, int8<\/td>\n<\/tr>\n<tr>\n<td>2024-01-08<\/td>\n<td><a href=\"https:\/\/huggingface.co\/Tele-AI\/telechat-7B\">Tele-AI\/telechat-7B<\/a><\/td>\n<td>text generation<\/td>\n<td>\u2014<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>int4, int8<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>From the Hugging Face API. Last changed 2026-09-28 (JST). Collected by code.<\/em><\/p>\n<h2>Licenses of Its Own Models (as recorded in our articles)<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>License<\/th>\n<th>Commercial use<\/th>\n<th>Models<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>apache-2.0<\/code><\/td>\n<td>Commercial use allowed<\/td>\n<td>XingChen-AGI\/Xing4.0-29B-A4B<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>The license can differ from model to model even within one publisher. This is not legal advice; check each model&#8217;s license text before use.<\/em><\/p>\n<h2>By Model Family<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Family<\/th>\n<th>Own-model articles<\/th>\n<th>Derived-model articles<\/th>\n<th>Smallest VRAM tier<\/th>\n<th>Latest<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/model-xingchen-agi-xing4-0-en\/\">Xing4.0<\/a><\/td>\n<td>1<\/td>\n<td>0<\/td>\n<td>24GB<\/td>\n<td>2026-09-28<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h2>Articles on Its Own Models<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Published<\/th>\n<th>Model<\/th>\n<th>Type<\/th>\n<th>Article<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>2026-09-28<\/td>\n<td>XingChen-AGI\/Xing4.0-29B-A4B<\/td>\n<td>New Models<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/28\/xing4-0-29b-a4b-china-telecom\/\">Xing4.0-29B-A4B 29B MoE Model Strong in Coding Agents: 24GB+ VRAM<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>Last updated 2026-09-28 (JST). The explanation at the top of this page was written with the help of AI from the primary sources it cites. The tables and lists under &#8220;Our Coverage and Data&#8221; are assembled by code from our article log and the Hugging Face API.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>About China Telecom (Xing, TeleChat) China Telecom is a major Chinese telecommunications carrier. Its AI units have released the TeleChat language models on Hugging Face. China Telecom&#8217;s Institute of Artificial Intelligence (TeleAI) published TeleChat on the Tele-AI organization (22), and since September 2026 its successor Xing has been released by China Telecom Artificial Intelligence Technology [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-6190","page","type-page","status-publish","hentry"],"lang":"en","translations":{"en":6190,"ja":6189},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages\/6190","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/comments?post=6190"}],"version-history":[{"count":0,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages\/6190\/revisions"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=6190"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}