docs: update parser launch examples

docs/configuration/server.md

@@ -109,7 +109,10 @@ about.
 | `--tool-call-parser` | Parser for OpenAI-compatible tool-call payloads (handled by the smg gateway). |
 | `--enable-custom-logit-processor` | Allow custom logit processors. Keep disabled unless the deployment needs it. |
 
-Common parser values include `kimi_k2` and `gpt-oss`.
+Common reasoning parser values include `kimi_k25`, `base`, `qwen3`, and
+`deepseek_r1`. Common tool-call parser values include `kimik2`, `qwen`, `json`,
+and `passthrough`. The parser names are validated by the SMG gateway, so use
+the values accepted by the bundled `tokenspeed-smg` package.
 
 ## Speculative Decoding
 

docs/guides/launching.md

@@ -31,8 +31,8 @@ tokenspeed serve nvidia/Kimi-K2.5-NVFP4 \
   --max-num-seqs 256 \
   --attention-backend trtllm_mla \
   --moe-backend flashinfer_trtllm \
-  --reasoning-parser kimi_k2 \
-  --tool-call-parser kimi_k2
+  --reasoning-parser kimi_k25 \
+  --tool-call-parser kimik2
 ```
 
 ## Launch Checklist

docs/recipes/models.md

@@ -24,8 +24,8 @@ tokenspeed serve nvidia/Kimi-K2.5-NVFP4 \
   --max-num-seqs 256 \
   --attention-backend trtllm_mla \
   --moe-backend flashinfer_trtllm \
-  --reasoning-parser kimi_k2 \
-  --tool-call-parser kimi_k2 \
+  --reasoning-parser kimi_k25 \
+  --tool-call-parser kimik2 \
   --host 0.0.0.0 \
   --port 8000
 ```
@@ -44,8 +44,7 @@ tokenspeed serve openai/gpt-oss-20b \
   --tensor-parallel-size 1 \
   --max-model-len 131072 \
   --chunked-prefill-size 8192 \
-  --reasoning-parser gpt-oss \
-  --tool-call-parser gpt-oss \
+  --reasoning-parser base \
   --host 0.0.0.0 \
   --port 8000
 ```
@@ -58,8 +57,7 @@ tokenspeed serve openai/gpt-oss-120b \
   --kv-cache-dtype fp8 \
   --chunked-prefill-size 8192 \
   --max-num-seqs 256 \
-  --reasoning-parser gpt-oss \
-  --tool-call-parser gpt-oss \
+  --reasoning-parser base \
   --host 0.0.0.0 \
   --port 8000
 ```

python/tokenspeed/runtime/utils/server_args.py

@@ -1312,7 +1312,7 @@ def add_cli_args(parser: argparse.ArgumentParser):
             type=str,
             default=ServerArgs.reasoning_parser,
             help=(
-                "Reasoning parser name (e.g. 'minimax', 'gpt-oss'). "
+                "Reasoning parser name (e.g. 'minimax', 'kimi_k25'). "
                 "Used to defer json_schema grammars past the model's "
                 "reasoning channel."
             ),