Document vLLM 0.25.1 production validation

README.md

@@ -57,6 +57,7 @@ The production profile uses:
 
 ```bash
 vllm serve WaveCut/gemma-4-26B-A4B-it-heretic-W4A16-G64-RTN-vllm \
+  --revision 51b3bbe6b47f12a1841e3934ce24e767bb8dec17 \
   --served-model-name "Gemma 4 26B Heretic" \
   --trust-remote-code \
   --dtype bfloat16 \
@@ -70,9 +71,18 @@ vllm serve WaveCut/gemma-4-26B-A4B-it-heretic-W4A16-G64-RTN-vllm \
   --tool-call-parser gemma4
 ```
 
-The historical RTX 3090 benchmark and harness are included under `bench/` and
-`scripts/`. A fresh vLLM 0.23.0 versus 0.25.1 text, vision, and long-context
-comparison will be added after the identical A/B run completes on AI Farm.
+The checkpoint was accepted for production on AI Farm with vLLM 0.25.1,
+Torch 2.11.0+cu130, Transformers 5.8.0, and compressed-tensors 0.17.0.
+On an RTX 3090, both vLLM 0.23.0 and 0.25.1 passed 13/13 text and vision case
+groups plus streaming checks. The matrix included exact needle recovery at
+39,186 text tokens and 38,490 multimodal tokens, including concurrency 2.
+Both runtimes also passed 24/24 production-shaped Russian quality probes;
+median probe latency improved from 0.6385 s to 0.6125 s on vLLM 0.25.1.
+
+The full environment, timing table, harness, and raw JSONL artifacts are in
+the [AI Farm vLLM report](https://github.com/iamwavecut/aifarm-vllm-openai/blob/af3a2915246a73585c175355af387c23e7360238/results/20260716-vllm-023-vs-0251.md).
+The historical RTX 3090 quantization benchmark and harness remain available
+under `bench/` and `scripts/`.
 
 ## Limitations