diff --git a/environments/benchmarks/tblite/local_vllm.yaml b/environments/benchmarks/tblite/local_vllm.yaml
index b6574a6bd..17689ba1d 100644
--- a/environments/benchmarks/tblite/local_vllm.yaml
+++ b/environments/benchmarks/tblite/local_vllm.yaml
@@ -4,7 +4,7 @@
 #
 # Start the vLLM server from the atropos directory:
 #   python -m example_trainer.vllm_api_server \
-#       --model Qwen/Qwen3-4B-Thinking-2507 \
+#       --model Qwen/Qwen3-4B-Instruct-2507 \
 #       --port 9001 \
 #       --gpu-memory-utilization 0.8 \
 #       --max-model-len=32000
@@ -26,14 +26,15 @@ env:
   task_timeout: 1200
   eval_concurrency: 8
   tool_call_parser: "hermes"
-  tokenizer_name: "Qwen/Qwen3-4B-Thinking-2507"
+  system_prompt: "You are an expert terminal agent. You MUST use the provided tools to complete tasks. Use the terminal tool to run shell commands, read_file to read files, write_file to write files, search_files to search, and patch to edit files. Do NOT write out solutions as text - execute them using the tools. Always start by exploring the environment with terminal commands."
+  tokenizer_name: "Qwen/Qwen3-4B-Instruct-2507"
   use_wandb: false
-  wandb_name: "tblite-qwen3-4b-thinking"
+  wandb_name: "tblite-qwen3-4b-instruct"
   ensure_scores_are_not_same: false
   data_dir_to_save_evals: "environments/benchmarks/evals/tblite-qwen3-4b-local"
 
 openai:
   base_url: "http://localhost:9001"
-  model_name: "Qwen/Qwen3-4B-Thinking-2507"
+  model_name: "Qwen/Qwen3-4B-Instruct-2507"
   server_type: "vllm"
   health_check: false