TeleChat2-3B
Tiny Apache-2.0 Chinese LLM from China Telecom for edge and Ascend NPUs.
About
TeleChat2-3B is a 3-billion-parameter open-weight language model from TeleAI, the AI division of China Telecom, released in July 2026 under Apache 2.0. It belongs to the TeleChat2 family, a multi-scale collection of Chinese-first LLMs developed with domestic infrastructure, including Huawei Ascend hardware.
The model is aimed at resource-constrained Chinese-language applications: conversational AI on consumer GPUs, a single RTX 4090, or Ascend Atlas edge NPUs. Its permissive license and small footprint make it interesting for telecom-domain assistants, customer-service bots, and government-adjacent deployments where data must stay on domestic stacks.
The family scales up to TeleChat2.5-115B for enterprise workloads, so the 3B tier is best understood as the edge and prototyping end of a coherent lineup. The model card on Modelers.cn is sparse, so capability claims should be validated by direct testing.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 104; Trending score 66/100 from researcher assessment.