Ollama v0.40.15 adds MiMo V2.6 for code agents and AutoTrust decision models (November 3, 2026)

What you get

Ollama v0.40.15, released on November 3, 2026, adds two new model families and a few fixes builders will notice.

• MiMo V2.6 (mimo-v2.6): Xiaomi MiMo's 9B model for code and agent workflows. It is fine-tuned from Qwen3.5-9B, reads text and images, and supports thinking and tool use. The context window is 256K. The download is 6.6GB to 8.7GB. It is released under the MIT license.
• AutoTrust decision models (autotrust-jev): they read text and return choices, yes or no probabilities and ratings for the questions you supply. Use them to route requests, classify documents or grade an answer against your own criteria. The family comes in 9b, 35b and 27b-vl sizes. The 35b and 27b-vl sizes can also read images. The default tag runs the 9B text model, an 11GB download.
• Other changes: ollama launch now supports Kimi Code CLI, the successor to kimi-cli, and offers to migrate kimi-cli setups. It also fills in Pi's model list, and Claude Desktop's model picker shows Ollama's real model names. Recurrent and vision GGUF models can now serve more than one request at a time. Running ollama with no arguments opens the welcome launcher.

Prerequisites

• Ollama v0.40.15 or later
• No API key. Local requests don't need one.
• Disk space for the model: about 6.6GB to 8.7GB for mimo-v2.6, 11GB for autotrust-jev

Steps

1. Update Ollama to v0.40.15.
2. Run MiMo V2.6:

ollama run mimo-v2.6

3. Or call it from Python, as shown on its library page:

from ollama import chat

response = chat(
    model='mimo-v2.6',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)