[Runtimes] Add GLM, Devstral, and MiniMax manifests

config/models/MiniMaxAI/MiniMax-M2.5.yaml

@@ -0,0 +1,24 @@
+apiVersion: ome.io/v1beta1
+kind: ClusterBaseModel
+metadata:
+  name: minimax-m2-5
+spec:
+  modelCapabilities:
+    - TEXT_TO_TEXT
+  vendor: MiniMaxAI
+  displayName: minimax.minimax-m2-5
+  modelArchitecture: MiniMaxM2ForCausalLM
+  disabled: false
+  version: "1.0.0"
+  modelFormat:
+    name: safetensors
+    version: "1.0.0"
+  modelFramework:
+    name: transformers
+    version: "4.46.1"
+  modelParameterSize: 229B
+  quantization: fp8
+  storage:
+    storageUri: hf://MiniMaxAI/MiniMax-M2.5
+    path: /raid/models/MiniMaxAI/MiniMax-M2.5
+    key: hf-token

config/models/mistralai/Devstral-2-123B-Instruct-2512.yaml

@@ -0,0 +1,24 @@
+apiVersion: ome.io/v1beta1
+kind: ClusterBaseModel
+metadata:
+  name: devstral-2-123b-instruct-2512
+spec:
+  modelCapabilities:
+    - TEXT_TO_TEXT
+  vendor: mistralai
+  displayName: mistralai.devstral-2-123b-instruct-2512
+  modelArchitecture: Ministral3ForCausalLM
+  disabled: false
+  version: "1.0.0"
+  modelFormat:
+    name: safetensors
+    version: "1.0.0"
+  modelFramework:
+    name: transformers
+    version: "5.0.0.dev0"
+  modelParameterSize: 123B
+  quantization: fp8
+  storage:
+    storageUri: hf://mistralai/Devstral-2-123B-Instruct-2512
+    path: /raid/models/mistralai/Devstral-2-123B-Instruct-2512
+    key: hf-token

config/models/mistralai/Devstral-Small-2-24B-Instruct-2512.yaml

@@ -0,0 +1,25 @@
+apiVersion: ome.io/v1beta1
+kind: ClusterBaseModel
+metadata:
+  name: devstral-small-2-24b-instruct-2512
+spec:
+  modelCapabilities:
+    - IMAGE_TEXT_TO_TEXT
+    - VIDEO_TEXT_TO_TEXT
+  vendor: mistralai
+  displayName: mistralai.devstral-small-2-24b-instruct-2512
+  modelArchitecture: Mistral3ForConditionalGeneration
+  disabled: false
+  version: "1.0.0"
+  modelFormat:
+    name: safetensors
+    version: "1.0.0"
+  modelFramework:
+    name: transformers
+    version: "5.0.0.dev0"
+  modelParameterSize: 24B
+  quantization: fp8
+  storage:
+    storageUri: hf://mistralai/Devstral-Small-2-24B-Instruct-2512
+    path: /raid/models/mistralai/Devstral-Small-2-24B-Instruct-2512
+    key: hf-token

config/models/zai-org/GLM-4.5V.yaml

@@ -8,8 +8,16 @@ spec:
     - VIDEO_TEXT_TO_TEXT
   vendor: zai-org
   displayName: zai-org.glm-4.5v
+  modelArchitecture: Glm4vMoeForConditionalGeneration
   disabled: false
   version: "1.0.0"
+  modelFormat:
+    name: safetensors
+    version: "1.0.0"
+  modelFramework:
+    name: transformers
+    version: "4.57.1"
+  modelParameterSize: 106B
   storage:
     storageUri: hf://zai-org/GLM-4.5V
     path: /raid/models/zai-org/GLM-4.5V

config/models/zai-org/GLM-4.6-FP8.yaml

@@ -0,0 +1,24 @@
+apiVersion: ome.io/v1beta1
+kind: ClusterBaseModel
+metadata:
+  name: glm-4-6-fp8
+spec:
+  modelCapabilities:
+    - TEXT_TO_TEXT
+  vendor: zai-org
+  displayName: zai-org.glm-4.6-fp8
+  modelArchitecture: Glm4MoeForCausalLM
+  disabled: false
+  version: "1.0.0"
+  modelFormat:
+    name: safetensors
+    version: "1.0.0"
+  modelFramework:
+    name: transformers
+    version: "4.54.0"
+  modelParameterSize: 355B
+  quantization: fp8
+  storage:
+    storageUri: hf://zai-org/GLM-4.6-FP8
+    path: /raid/models/zai-org/GLM-4.6-FP8
+    key: hf-token

config/models/zai-org/GLM-4.7-FP8.yaml

@@ -0,0 +1,24 @@
+apiVersion: ome.io/v1beta1
+kind: ClusterBaseModel
+metadata:
+  name: glm-4-7-fp8
+spec:
+  modelCapabilities:
+    - TEXT_TO_TEXT
+  vendor: zai-org
+  displayName: zai-org.glm-4.7-fp8
+  modelArchitecture: Glm4MoeForCausalLM
+  disabled: false
+  version: "1.0.0"
+  modelFormat:
+    name: safetensors
+    version: "1.0.0"
+  modelFramework:
+    name: transformers
+    version: "4.52.3"
+  modelParameterSize: 355B
+  quantization: fp8
+  storage:
+    storageUri: hf://zai-org/GLM-4.7-FP8
+    path: /raid/models/zai-org/GLM-4.7-FP8
+    key: hf-token

config/runtimes/kustomization.yaml

@@ -32,6 +32,7 @@ resources:
 - srt/mistral-7b-instruct-rt.yaml
 - srt/mixtral-8x7b-instruct-pd-rt.yaml
 - srt/mixtral-8x7b-instruct-rt.yaml
+- srt/zai-org/glm-text-fp8-tp8-rt.yaml
 # vLLM runtimes
 - vllm/e5-mistral-7b-instruct-rt.yaml
 - vllm/llama-3-1-405b-instruct-fp8-rt.yaml
@@ -50,3 +51,7 @@ resources:
 - vllm/llama-4-scout-17b-16e-instruct-rt.yaml
 - vllm/mistral-7b-instruct-rt.yaml
 - vllm/mixtral-8x7b-instruct-rt.yaml
+- vllm/zai-org/glm-vl-tp4-rt.yaml
+- vllm/minimax/minimax-m2-fp8-tp8-rt.yaml
+- vllm/mistralai/devstral-2-123b-instruct-fp8-tp4-rt.yaml
+- vllm/mistralai/devstral-small-2-24b-instruct-tp2-rt.yaml