Repository navigation
models glm-5.2, glm-5.1, kimi-k2.7-code, deepseek-v4-flash fail #129
Description
Activity
- changed the title
[-]models glm-5.2, glm-5.1, kimi-k2.7, kimi-k2.6, deepseek-v4-pro, deepseek-v4-flash fail[/-][+]models glm-5.2, glm-5.1, kimi-k2.7-code, deepseek-v4-flash fail[/+]on Jun 21, 2026 Merge-state update checked on 2026-10-06: PR #748 has merged and replaces the retired NVIDIA Build GLM-5.2 default with GLM-5.3. Earlier PR #279 added endpoint metadata.
Keeping this issue open. Updating the default and metadata does not establish successful inference for every reported model or resolve the subsequent mimo-v2.5/minimax-m3 deadlock reports. This audit did not run live provider acceptance tests.
and mimo-v2.5 looks to be in a deadlock https://github.com/generic-automation-and-it/smooth-devex-template/actions/runs/27897636303/job/82553097454?pr=53#logs
minimax-m3 https://github.com/generic-automation-and-it/smooth-devex-template/actions/runs/27897478044/job/82551587029
deepseek https://github.com/generic-automation-and-it/smooth-devex-template/actions/runs/27897636303/job/82552018184?pr=53
and there is more... glm-5.2, kimi-k2.7-code, deepseek-v4-flash fail
maybe make the skripts more llm resistant or recommend open source models for testing