feat: add Qwen3.8 Flash Next renderer

renderers/base.py

@@ -1033,6 +1033,7 @@ def bridge_to_next_turn(self, *args: Any, **kwargs: Any) -> "RenderedTokens | No
     "Qwen/Qwen3.6-35B-A3B": "qwen3.6",
     # Qwen3.8.
     "Qwen/Qwen3.8-27B": "qwen3.8",
+    "Qwen/Qwen3.8-Flash-Next": "qwen3.8",
     # Qwen3-VL.
     "Qwen/Qwen3-VL-4B-Instruct": "qwen3-vl",
     "Qwen/Qwen3-VL-8B-Instruct": "qwen3-vl",
@@ -1140,6 +1141,7 @@ def bridge_to_next_turn(self, *args: Any, **kwargs: Any) -> "RenderedTokens | No
     "Qwen/Qwen3.6-35B-A3B": {"image"},
     # Qwen3.8 adds reasoning-effort control and preserves thinking by default.
     "Qwen/Qwen3.8-27B": {"image"},
+    "Qwen/Qwen3.8-Flash-Next": {"image"},
     # Kimi K2.5 / K2.6 are unified VLMs (HF tag ``image-text-to-text``)
     # with custom processor (``KimiK25Processor`` + ``KimiK25VisionProcessor``).
     # Vision wrap is different from Qwen-VL:

renderers/qwen35.py

@@ -104,6 +104,7 @@
     "Qwen/Qwen3.6-35B-A3B": True,
     # Qwen3.8 keeps thinking enabled by default.
     "Qwen/Qwen3.8-27B": True,
+    "Qwen/Qwen3.8-Flash-Next": True,
 }
 
 

tests/conftest.py

@@ -23,6 +23,7 @@
     ("Qwen/Qwen3.5-9B", "auto"),
     ("Qwen/Qwen3.6-35B-A3B", "auto"),
     ("Qwen/Qwen3.8-27B", "auto"),
+    ("Qwen/Qwen3.8-Flash-Next", "auto"),
     ("Qwen/Qwen3-VL-4B-Instruct", "auto"),
     ("google/gemma-4-31B-it", "auto"),
     ("zai-org/GLM-5", "auto"),

tests/test_bridge.py

@@ -27,6 +27,7 @@
     ("Qwen/Qwen3.5-9B", "auto"),
     ("Qwen/Qwen3.6-35B-A3B", "auto"),
     ("Qwen/Qwen3.8-27B", "auto"),
+    ("Qwen/Qwen3.8-Flash-Next", "auto"),
     ("google/gemma-4-31B-it", "auto"),
     ("thinkingmachines/Inkling", "auto"),
     ("thinkingmachines/Inkling-Small", "auto"),

tests/test_client.py

@@ -385,8 +385,9 @@ def test_generate_threads_prompt_attribution_through_prebuilt_prompt_path():
         ("Qwen/Qwen3-VL-4B-Instruct", "renderers.qwen3_vl:Qwen3VLRenderer"),
         ("Qwen/Qwen3.5-2B", "renderers.qwen35:Qwen35Renderer"),
         ("Qwen/Qwen3.8-27B", "renderers.qwen38:Qwen38Renderer"),
+        ("Qwen/Qwen3.8-Flash-Next", "renderers.qwen38:Qwen38Renderer"),
     ],
-    ids=["qwen3_vl", "qwen35", "qwen38"],
+    ids=["qwen3_vl", "qwen35", "qwen38", "qwen38_flash_next"],
 )
 def test_generate_serializes_multimodal_features_for_qwen_vl_family(
     model_id, renderer_class_path

tests/test_disabled_thinking_stability.py

@@ -45,6 +45,10 @@
         "Qwen/Qwen3.8-27B",
         Qwen38RendererConfig(enable_thinking=False),
     ),
+    (
+        "Qwen/Qwen3.8-Flash-Next",
+        Qwen38RendererConfig(enable_thinking=False),
+    ),
     (
         "google/gemma-4-26B-A4B-it",
         Gemma4RendererConfig(enable_thinking=False),

tests/test_qwen38.py

@@ -31,10 +31,12 @@
     }
 ]
 
+QWEN38_MODELS = ["Qwen/Qwen3.8-27B", "Qwen/Qwen3.8-Flash-Next"]
 
-@lru_cache(maxsize=1)
-def _qwen38():
-    tokenizer = load_tokenizer("Qwen/Qwen3.8-27B")
+
+@lru_cache(maxsize=None)
+def _qwen38(model_name):
+    tokenizer = load_tokenizer(model_name)
     return tokenizer, create_renderer(tokenizer)
 
 
@@ -48,10 +50,11 @@ def _expected(tokenizer, messages, **kwargs):
     return list(result)
 
 
-def test_qwen38_is_registered_with_native_defaults():
-    tokenizer, renderer = _qwen38()
+@pytest.mark.parametrize("qwen38_model", QWEN38_MODELS)
+def test_qwen38_is_registered_with_native_defaults(qwen38_model):
+    tokenizer, renderer = _qwen38(qwen38_model)
 
-    assert tokenizer.name_or_path == "Qwen/Qwen3.8-27B"
+    assert tokenizer.name_or_path == qwen38_model
     assert MODEL_RENDERER_MAP[tokenizer.name_or_path] == "qwen3.8"
     assert MULTIMODAL_MODELS[tokenizer.name_or_path] == {"image"}
     assert isinstance(renderer, Qwen38Renderer)
@@ -86,8 +89,9 @@ def test_qwen38_config_discriminator():
         pytest.param({"preserve_thinking": False}, id="drop-history-thinking"),
     ],
 )
-def test_qwen38_text_and_tool_parity(config_kwargs):
-    tokenizer, _ = _qwen38()
+@pytest.mark.parametrize("qwen38_model", QWEN38_MODELS)
+def test_qwen38_text_and_tool_parity(config_kwargs, qwen38_model):
+    tokenizer, _ = _qwen38(qwen38_model)
     renderer = Qwen38Renderer(tokenizer, Qwen38RendererConfig(**config_kwargs))
     cases = [
         (
@@ -152,8 +156,9 @@ def test_qwen38_text_and_tool_parity(config_kwargs):
         assert renderer.render_ids(messages, **render_kwargs) == expected
 
 
-def test_qwen38_keeps_inline_think_markup_in_visible_content():
-    tokenizer, renderer = _qwen38()
+@pytest.mark.parametrize("qwen38_model", QWEN38_MODELS)
+def test_qwen38_keeps_inline_think_markup_in_visible_content(qwen38_model):
+    tokenizer, renderer = _qwen38(qwen38_model)
     messages = [
         {"role": "user", "content": "Echo this."},
         {
@@ -165,8 +170,9 @@ def test_qwen38_keeps_inline_think_markup_in_visible_content():
     assert renderer.render_ids(messages) == _expected(tokenizer, messages)
 
 
-def test_qwen38_requires_a_real_user_query():
-    _, renderer = _qwen38()
+@pytest.mark.parametrize("qwen38_model", QWEN38_MODELS)
+def test_qwen38_requires_a_real_user_query(qwen38_model):
+    _, renderer = _qwen38(qwen38_model)
 
     with pytest.raises(ValueError, match="No user query found"):
         renderer.render_ids([{"role": "system", "content": "System only."}])

tests/test_renderer_config_parity.py

@@ -48,6 +48,7 @@
     ("Qwen/Qwen3.5-9B", "auto"),
     ("Qwen/Qwen3.6-35B-A3B", "auto"),
     ("Qwen/Qwen3.8-27B", "auto"),
+    ("Qwen/Qwen3.8-Flash-Next", "auto"),
     ("google/gemma-4-31B-it", "auto"),
     ("zai-org/GLM-5", "auto"),
     ("zai-org/GLM-5.1", "auto"),

tests/test_roundtrip.py

@@ -33,6 +33,7 @@
     ("Qwen/Qwen3.5-9B", "auto"),
     ("Qwen/Qwen3.6-35B-A3B", "auto"),
     ("Qwen/Qwen3.8-27B", "auto"),
+    ("Qwen/Qwen3.8-Flash-Next", "auto"),
     ("Qwen/Qwen3-VL-4B-Instruct", "auto"),
     ("google/gemma-4-31B-it", "auto"),
     ("thinkingmachines/Inkling", "auto"),