Ollama v0.40.16: /v1/responses stops cutting long input silently (November 9, 2026)
Ollama v0.40.16: /v1/responses follows truncation and no longer cuts input silently, and /v1/models returns [] instead of null.
3 of 3 free copies left this month.Unlimited with Pro
What changed
Ollama v0.40.16 (released November 9) changes two of its OpenAI-compatible endpoints:
/v1/responsesrespectstruncationand no longer silently truncates input by default./v1/modelsreturns an empty list instead ofnullwhen no models are installed.
The release also adds Ripple Preview, a 65B dense model for agentic coding (ollama run ripple-preview), and Qwen3.8 Flash Next uses about 27 GiB less memory on Apple silicon.
Prerequisites
- Ollama v0.40.16 or later, running locally
- No key is needed for local models
Steps
- Upgrade Ollama to v0.40.16 or later.
- If your code sends long input to
/v1/responsesand relied on Ollama cutting it, settruncationin the request yourself. - If your code reads
/v1/models, make sure it handles an empty list.
Test it
On a machine with no models installed, call /v1/models: the data list should be empty, not null. Then send a /v1/responses request longer than the model's context and check what comes back.
Not confirmed
The release notes do not say what /v1/responses now returns when input is too long and truncation is not set.
