pypi / modelscope

modelscope

What changed in modelscope 1, from its official release notes, upgrade guides and advisories. Sources checked 2026-10-24.

1 breaking8 new APIs and options2 behavior and defaults2 tooling1 fixes and other

See which of these apply to your code, ranked, in your agent:

npx @unstale/cli init

modelscope 114 changes

The modelscope and ms console scripts are now provided solely by modelscope-hub, which is required at version 0.3.0 or later.
1.40.0release notes

Intel XPU, including iGPU and discrete GPUs such as Arc, Flex, and Max, is supported as a target device for model inference and training with IPEX-LLM optimizations.
1.40.2release notes
SDK and CLI compatibility shims add agent identity management, including key rotation, status, and signed token issuance.
1.40.1release notes
MCPApi adds a method to deploy an MCP server.
1.40.1release notes
A speech separation model for 2-speaker 8 kHz audio is available.
1.40.1release notes
A speech separation model for 2-speaker 8 kHz audio is available.
1.40.1release notes
A native inference pipeline for Joint Acoustic Echo Cancellation is available under the audio/AEC module.
1.40.1release notes
The Ascend Docker image supports building mcore-bridge from source with a configurable branch or version for Megatron-based training environments.
1.40.1release notes
MCPApi adds a method to undeploy an MCP server.
1.40.1release notes

MCP discovery, operational-server listing, and server-detail requests now go through modelscope-hub for consistent authentication, retry, and error handling.
1.40.1release notes
Upload environment variables are unified, and custom README content is added for uploading AIGC model weights.
1.40.1release notes

Docker images can now be built for minor version releases, improving release granularity for containerized deployments.
1.40.2release notes
Studio CLI is consolidated as a re-export of the hub implementation.
1.40.1release notes