{#-
qwen38_flexible.jinja
=====================================================================
Drop-in replacement chat template for Qwen/Qwen3.8-27B (and other
Qwen3.8 models), forked from the upstream chat_template.jinja.
Same wire format (<|im_start|>/<|im_end|>, blocks, tool_call
XML), but without the rigid conversation-shape enforcement.
RELAXED vs upstream:
R1 No 'No user query found in messages.' raise. Assistant-first,
user-less, or tool-response-only transcripts render fine.
R2 System messages are allowed ANYWHERE in the transcript, not
just at index 0. Mid-conversation system notes render as their
own <|im_start|>system block instead of raising
'System message must be at the beginning.'
R3 'developer' role accepted as an alias for 'system' (OpenAI
Responses-API convention used by modern harnesses).
R4 Unknown roles (narrator, char, example_user, ...) render
verbatim as <|im_start|>{role} ... <|im_end|> instead of
raising 'Unexpected message role.'.
R5 continue_final_message support: when true and the last message
is an assistant turn, that turn is left open (no <|im_end|>)
for prefilling, and no generation prompt is appended.
R6 No alternation assumptions anywhere: consecutive user turns,
consecutive assistant turns, assistant-leading transcripts and
trailing assistant turns all render as-is.
KEPT from upstream (byte-compatible rendering):
- render_content macro with vision counters
- reasoning_effort (xhigh / medium / low) instruction injection
- preserve_thinking (defaults TRUE on 3.8) + last_query_index walk
- tool definitions block and / rendering
- generation prompt with enable_thinking gate
USAGE
llama.cpp : llama-server -m qwen38-27b-*.gguf \
--chat-template-file qwen38_flexible.jinja
vLLM : vllm serve Qwen/Qwen3.8-27B \
--chat-template qwen38_flexible.jinja \
--tool-call-parser qwen3_coder --reasoning-parser qwen3
SGLang : python -m sglang.launch_server --model-path Qwen/Qwen3.8-27B \
--chat-template qwen38_flexible.jinja \
--tool-call-parser qwen3_coder --reasoning-parser qwen3
Forked from Qwen/Qwen3.8-27B/chat_template.jinja (2026-08-13).
License: Apache 2.0 (same as upstream).
-#}
{%- set image_count = namespace(value=0) %}
{%- set video_count = namespace(value=0) %}
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
{%- if content is string %}
{{- content }}
{%- elif content is iterable and content is not mapping %}
{%- for item in content %}
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
{%- if is_system_content %}
{{- raise_exception('System message cannot contain images.') }}
{%- endif %}
{%- if do_vision_count %}
{%- set image_count.value = image_count.value + 1 %}
{%- endif %}
{%- if add_vision_id is defined and add_vision_id %}
{{- 'Picture ' ~ image_count.value ~ ': ' }}
{%- endif %}
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
{%- elif 'video' in item or item.type == 'video' %}
{%- if is_system_content %}
{{- raise_exception('System message cannot contain videos.') }}
{%- endif %}
{%- if do_vision_count %}
{%- set video_count.value = video_count.value + 1 %}
{%- endif %}
{%- if add_vision_id is defined and add_vision_id %}
{{- 'Video ' ~ video_count.value ~ ': ' }}
{%- endif %}
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
{%- elif 'text' in item %}
{{- item.text }}
{%- else %}
{{- raise_exception('Unexpected item type in content.') }}
{%- endif %}
{%- endfor %}
{%- elif content is none or content is undefined %}
{{- '' }}
{%- else %}
{{- raise_exception('Unexpected content type.') }}
{%- endif %}
{%- endmacro %}
{%- if not messages %}
{{- raise_exception('No messages provided.') }}
{%- endif %}
{#- reasoning_effort instruction injection (identical to upstream) -#}
{%- set reasoning_instructions = '' %}
{%- if enable_thinking is undefined or enable_thinking is true %}
{%- set resolved_reasoning_effort = reasoning_effort|default('xhigh') %}
{%- if resolved_reasoning_effort not in ('xhigh', 'medium', 'low') %}
{{- raise_exception('Unexpected reasoning effort ' ~ reasoning_effort ~ '. Supported types are xhigh (default), medium, and low.') }}
{%- endif %}
{%- if resolved_reasoning_effort == 'xhigh' %}
{%- set reasoning_instructions = 'Reasoning effort is set to xhigh. Please think carefully through the task, validate key assumptions, consider plausible alternatives, and prioritize correctness, consistency, and clarity in the final answer.' %}
{%- elif resolved_reasoning_effort == 'low' %}
{%- set reasoning_instructions = 'Reasoning effort is set to low. Keep your thinking brief and focused, moving directly to the conclusion without unnecessary elaboration.' %}
{%- endif %}
{%- endif %}
{#- R3: 'developer' at index 0 is treated exactly like 'system' -#}
{%- set first_is_system = messages[0].role == 'system' or messages[0].role == 'developer' %}
{%- if tools and tools is iterable and tools is not mapping %}
{{- '<|im_start|>system\n' }}
{%- if reasoning_instructions %}
{{- reasoning_instructions + '\n\n' }}
{%- endif %}
{{- "# Tools\n\nYou have access to the following functions:\n\n" }}
{%- for tool in tools %}
{{- "\n" }}
{%- if tool.function is defined %}
{{- tool.function | tojson }}
{%- else %}
{{- tool | tojson }}
{%- endif %}
{%- endfor %}
{{- "\n" }}
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n\n\n\nvalue_1\n\n\nThis is the value for the second parameter\nthat can span\nmultiple lines\n\n\n\n\n\nReminder:\n- Function calls MUST follow the specified format: an inner block must be nested within XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n' }}
{%- if first_is_system %}
{%- set content = render_content(messages[0].content, false, true)|trim %}
{%- if content %}
{{- '\n\n' + content }}
{%- endif %}
{%- endif %}
{{- '<|im_end|>\n' }}
{%- else %}
{%- if first_is_system %}
{%- set content = render_content(messages[0].content, false, true)|trim %}
{%- if content %}
{{- '<|im_start|>system\n' + (reasoning_instructions + '\n\n' if reasoning_instructions else '') + content + '<|im_end|>\n' }}
{%- elif reasoning_instructions %}
{{- '<|im_start|>system\n' + reasoning_instructions + '<|im_end|>\n' }}
{%- endif %}
{%- elif reasoning_instructions %}
{{- '<|im_start|>system\n' + reasoning_instructions + '<|im_end|>\n' }}
{%- endif %}
{%- endif %}
{#- last_query_index walk (identical to upstream logic) -#}
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
{%- for message in messages[::-1] %}
{%- set index = (messages|length - 1) - loop.index0 %}
{%- if ns.multi_step_tool and message.role == "user" %}
{%- set content = render_content(message.content, false)|trim %}
{%- if not(content.startswith('') and content.endswith('')) %}
{%- set ns.multi_step_tool = false %}
{%- set ns.last_query_index = index %}
{%- endif %}
{%- endif %}
{%- endfor %}
{#- R1: upstream raised 'No user query found in messages.' here when the
transcript has no non-tool-response user turn (assistant-first, RP
transcripts, pure continuations). We simply keep the default index:
with preserve_thinking=true (3.8 default) the index is not consulted
anyway, and with preserve_thinking=false it only affects which
historical blocks survive. -#}
{%- if ns.multi_step_tool %}
{%- set ns.last_query_index = messages|length - 1 %}
{%- endif %}
{%- for message in messages %}
{%- set content = render_content(message.content, true)|trim %}
{%- if message.role == "system" or message.role == "developer" %}
{#- R2/R3: index 0 was already merged into the header above. A
system/developer note LATER in the transcript renders as its
own block instead of raising. -#}
{%- if not loop.first %}
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
{%- endif %}
{%- elif message.role == "user" %}
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
{%- elif message.role == "assistant" %}
{%- set reasoning_content = '' %}
{%- if message.reasoning_content is string %}
{%- set reasoning_content = message.reasoning_content %}
{%- else %}
{#- Rescue thinking embedded in content (harnesses that stuff
into the content string). Handles and
whitespace variants emitted by some quant runtimes. -#}
{%- set think_end = '' %}
{%- if '' in content %}
{%- set think_end = '' %}
{%- elif '' in content %}
{%- set think_end = '' %}
{%- elif '' in content %}
{%- set think_end = '' %}
{%- elif '' in content %}
{%- set think_end = '' %}
{%- endif %}
{%- if think_end %}
{%- set parts = content.split(think_end) %}
{%- set reasoning_content = parts[0] %}
{%- set content = parts[1:] | join(think_end) %}
{%- if '' in reasoning_content %}
{%- set reasoning_content = reasoning_content.split('')[1:] | join('') %}
{%- endif %}
{%- endif %}
{%- endif %}
{%- set reasoning_content = reasoning_content | trim %}
{%- set content = content | trim %}
{%- if preserve_thinking is undefined or preserve_thinking is true or loop.index0 > ns.last_query_index %}
{{- '<|im_start|>' + message.role + '\n\n' + reasoning_content + '\n\n\n' + content }}
{%- else %}
{{- '<|im_start|>' + message.role + '\n' + content }}
{%- endif %}
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
{%- if '' in content %}
{%- set content = content.split('')[0] | trim %}
{%- endif %}
{%- for tool_call in message.tool_calls %}
{%- if tool_call.function is defined %}
{%- set tool_call = tool_call.function %}
{%- endif %}
{%- if loop.first %}
{%- if content|trim %}
{{- '\n\n\n\n' }}
{%- else %}
{{- '\n\n' }}
{%- endif %}
{%- else %}
{{- '\n\n\n' }}
{%- endif %}
{%- if tool_call.arguments is defined and tool_call.arguments != '' %}
{%- if tool_call.arguments is mapping %}
{%- for args_name, args_value in tool_call.arguments|items %}
{{- '\n' }}
{%- set args_value = args_value | string if args_value is string else args_value | tojson | safe %}
{{- args_value }}
{{- '\n\n' }}
{%- endfor %}
{%- elif tool_call.arguments is string %}
{{- raise_exception("qwen38_flexible: tool_call.arguments must be a JSON object (mapping), got a string. Deserialize the JSON-encoded arguments once on ingest and store the resulting object.") }}
{%- endif %}
{%- endif %}
{{- '\n' }}
{%- endfor %}
{%- endif %}
{#- R5: leave the final assistant turn OPEN (no <|im_end|>) when
prefilling via continue_final_message. -#}
{%- if continue_final_message is defined and continue_final_message and loop.last %}
{%- else %}
{{- '<|im_end|>\n' }}
{%- endif %}
{%- elif message.role == "tool" %}
{%- if loop.previtem and loop.previtem.role != "tool" %}
{{- '<|im_start|>user' }}
{%- endif %}
{{- '\n\n' }}
{{- content }}
{{- '\n' }}
{%- if not loop.last and loop.nextitem.role != "tool" %}
{{- '<|im_end|>\n' }}
{%- elif loop.last %}
{{- '<|im_end|>\n' }}
{%- endif %}
{%- else %}
{#- R4: unknown roles render verbatim instead of raising.
Handy for narrators / multi-character / example transcripts. -#}
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
{%- endif %}
{%- endfor %}
{%- if add_generation_prompt and not (continue_final_message is defined and continue_final_message) %}
{{- '<|im_start|>assistant\n' }}
{%- if enable_thinking is defined and enable_thinking is false %}
{{- '\n\n\n\n' }}
{%- else %}
{{- '\n' }}
{%- endif %}
{%- endif %}