Update README.md

README.md

@@ -112,7 +112,7 @@ The model version is FP4 QAD 1.0 and is quantized with nvidia-modelopt **v0.47.0
 **Datasets:** GPQA Diamond, ToolEvalBench <br>
 **Data Collection Method by dataset:** Hybrid: Automated, Manually-Collected <br>
 **Labeling Method by dataset:** Hybrid: Manually-Labeled, Automated <br>
-**Properties:** GPQA Diamond contains 448 graduate-level multiple-choice questions written by domain experts in biology, physics, and chemistry. ToolEvalBench is a tool-calling quality benchmark that runs deterministic multi-turn agent scenarios against mock tools, scoring tool selection, parameter precision, multi-step chains, error recovery, and structured output. <br>
+**Properties:** GPQA Diamond contains 198 graduate-level multiple-choice questions written by domain experts in biology, physics, and chemistry. ToolEvalBench is a tool-calling quality benchmark that runs deterministic multi-turn agent scenarios against mock tools, scoring tool selection, parameter precision, multi-step chains, error recovery, and structured output. <br>
 
 The accuracy benchmark results are presented in the table below:
 
@@ -227,7 +227,7 @@ vllm serve nvidia/Qwen3.8-Flash-Next-FP4-Local \
 
 ### NVIDIA RTX PRO 6000 Blackwell
 
-vLLM container: [`ghcr.io/local-inference-lab/vllm:karmic-kraken`](https://github.com/local-inference-lab/blackwell-llm-docker/pkgs/container/vllm)
+vLLM container: [`ghcr.io/local-inference-lab/vllm:release-10142026`](https://github.com/local-inference-lab/blackwell-llm-docker/pkgs/container/vllm)
 
 Start the container. Set `HF_TOKEN` in the host environment first.
 
@@ -242,7 +242,7 @@ docker run --rm -it \
   -e HF_TOKEN=$HF_TOKEN \
   -v ~/.cache/huggingface:/root/.cache/huggingface \
   --entrypoint /bin/bash \
-  ghcr.io/local-inference-lab/vllm:karmic-kraken
+  ghcr.io/local-inference-lab/vllm:release-10142026
 ```
 
 Then launch vLLM inside the container: