[Test] E2E Nemotron-3-Super tests (#36803)

Signed-off-by: Roi Koren <roik@nvidia.com>

.buildkite/test_areas/lm_eval.yaml

@@ -90,6 +90,7 @@ steps:
   commands:
     - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/moe-refactor-dp-ep/config-b200.txt
 
+
 - label: GPQA Eval (GPT-OSS) (H100)
   timeout_in_minutes: 120
   device: h100

tests/evals/gsm8k/configs/Nemotron-3-Super-120B-A12B-BF16.yaml

+model_name: "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16"
+accuracy_threshold: 0.93
+num_questions: 1319
+num_fewshot: 5
+startup_max_wait_seconds: 1200
+server_args: >-
+  --enforce-eager
+  --max-model-len 4096
+  --tensor-parallel-size 8
+  --enable-expert-parallel
+  --speculative-config '{"method":"mtp","num_speculative_tokens":5}'

tests/evals/gsm8k/configs/Nemotron-3-Super-120B-A12B-FP8.yaml

+model_name: "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8"
+accuracy_threshold: 0.93
+num_questions: 1319
+num_fewshot: 5
+startup_max_wait_seconds: 1200
+server_args: >-
+  --enforce-eager
+  --max-model-len 4096
+  --tensor-parallel-size 8
+  --enable-expert-parallel
+  --speculative-config '{"method":"mtp","num_speculative_tokens":5}'

tests/evals/gsm8k/configs/Nemotron-3-Super-120B-A12B-NVFP4.yaml

+model_name: "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4"
+accuracy_threshold: 0.93
+num_questions: 1319
+num_fewshot: 5
+startup_max_wait_seconds: 1200
+server_args: >-
+  --enforce-eager
+  --max-model-len 4096
+  --tensor-parallel-size 2
+  --enable-expert-parallel
+  --speculative-config '{"method":"mtp","num_speculative_tokens":5}'

tests/evals/gsm8k/configs/models-blackwell.txt

@@ -5,3 +5,4 @@ DeepSeek-V2-Lite-Instruct-FP8.yaml
 Qwen3-30B-A3B-NVFP4.yaml
 Qwen3-Next-80B-A3B-NVFP4-EP2.yaml
 Qwen3-Next-FP8-EP2.yaml
+Nemotron-3-Super-120B-A12B-NVFP4.yaml

tests/evals/gsm8k/configs/models-h200.txt

@@ -2,3 +2,5 @@ DeepSeek-R1-TP.yaml
 DeepSeek-R1-DP.yaml
 DeepSeek-V3.2-TP.yaml
 DeepSeek-V3.2-DP.yaml
+Nemotron-3-Super-120B-A12B-BF16.yaml
+Nemotron-3-Super-120B-A12B-FP8.yaml