Eval bug: Unable to generate parser for this template, System message must be at the beginnin

[oleg-kn]

Name and Version

version: 7803 (b70d251)

Operating systems

Linux

GGML backends

CUDA

Hardware

2x RTX 3090

Models

unsloth/Qwen3.5-27B-GGUF:Q8_K_XL

Problem description & steps to reproduce

using Codex v0.94.0 i receive such message:

{"error":{"code":400,"message":"Unable to generate parser for this template. Automatic parser generation failed: \n------------\nWhile executing CallExpression at line 85, column 32
in source:\n...first %}↵            {{- raise_exception('System message must be at the beginnin...\n                                           ^\nError: Jinja Exception: System message
must be at the beginning.","type":"invalid_request_error"}}

First Bad Commit

No response

Relevant log output

srv operator(): got exception: {"error":{"code":400,"message":"Unable to generate parser for this template. Automatic parser generation failed: \n------------\nWhile executing CallExpression at line 85, column 32 in source:\n...first %}↵ {{- raise_exception('System message must be at the beginnin...\n ^\nError: Jinja Exception: System message must be at the beginning.","type":"invalid_request_error"}}
srv log_server_r: done request: POST /chat/completions

[Vivida1]

Same issue here using latest llama.cpp:server-cuda image.

llama-server -hf unsloth/Qwen3.5-35B-A3B-GGUF:Q8_0 --port 8080 --host 0.0.0.0 -c 0 --temp 0.6 --top-p 0.95 --top-k 20 --min-p 0.0 --chat-template-kwargs '{"enable_thinking":true}'

[org-tekeli-borisp]

Same issue here with latest llama.cpp (48cda24) compiled with
cmake -B build -DGGML_CUDA=ON && cmake --build build --config Release on Arch Linux.

llama-cli --version
ggml_cuda_init: found 3 CUDA devices (Total VRAM: 44738 MiB):
  Device 0: Tesla T4, compute capability 7.5, VMM: yes, VRAM: 14912 MiB
  Device 1: Tesla T4, compute capability 7.5, VMM: yes, VRAM: 14912 MiB
  Device 2: Tesla T4, compute capability 7.5, VMM: yes, VRAM: 14912 MiB
version: 8555 (48cda24c1)
built with GNU 15.2.1 for Linux x86_64

Running with
GGML_CUDA_ENABLE_UNIFIED_MEMORY=1 Development/src/llama.cpp/build/bin/llama-server -hf unsloth/Qwen3.5-27B-GGUF:Q8_0 --ctx-size 262144 --threads 12 --threads-batch 12 --tensor-split 0.33,0.33,0.34 --n-gpu-layers 2000000 --flash-attn on --fit on --batch-size 4096 --ubatch-size 1024 --host 0.0.0.0 --kv-unified --reasoning off --parallel 1 --jinja

[JohnGalt1717]

I'm having this too using the openai endpoints. Using openai itself works fine.

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.

[JohnGalt1717]

Please reopen. This is a major incompatibility with OpenAI’s api.

[chenxqiyu]

codex.jinja

{%- 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 %}
                    {{- '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 %}
                    {{- '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 %}
{%- if tools and tools is iterable and tools is not mapping %}
    {{- '<|im_start|>system\n' }}
    {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
    {%- for tool in tools %}
        {{- "\n" }}
        {{- tool | tojson }}
    {%- endfor %}
    {{- "\n</tools>" }}
    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> 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</IMPORTANT>' }}
    {%- if messages[0].role == 'system' %}
        {%- set content = render_content(messages[0].content, false, true)|trim %}
        {%- if content %}
            {{- '\n\n' + content }}
        {%- endif %}
    {%- endif %}
    {{- '<|im_end|>\n' }}
{%- else %}
    {%- if messages[0].role == 'system' %}
        {%- set content = render_content(messages[0].content, false, true)|trim %}
        {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
    {%- endif %}
{%- endif %}
{%- 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('<tool_response>') and content.endswith('</tool_response>')) %}
            {%- set ns.multi_step_tool = false %}
            {%- set ns.last_query_index = index %}
        {%- endif %}
    {%- endif %}
{%- endfor %}
{%- if ns.multi_step_tool %}
    {{- raise_exception('No user query found in messages.') }}
{%- endif %}
{%- for message in messages %}
    {%- set content = render_content(message.content, true)|trim %}
    {%- if message.role == "system" %}
        {%- if not loop.first %}
            {{- '<|im_start|>user\n[System Instruction]\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 %}
            {%- if '</think>' in content %}
                {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
                {%- set content = content.split('</think>')[-1].lstrip('\n') %}
            {%- endif %}
        {%- endif %}
        {%- set reasoning_content = reasoning_content|trim %}
        {%- if loop.index0 > ns.last_query_index %}
            {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\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 %}
            {%- 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<tool_call>\n<function=' + tool_call.name + '>\n' }}
                    {%- else %}
                        {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
                    {%- endif %}
                {%- else %}
                    {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
                {%- endif %}
                {%- if tool_call.arguments is defined %}
                    {%- for args_name, args_value in tool_call.arguments|items %}
                        {{- '<parameter=' + args_name + '>\n' }}
                        {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
                        {{- args_value }}
                        {{- '\n</parameter>\n' }}
                    {%- endfor %}
                {%- endif %}
                {{- '</function>\n</tool_call>' }}
            {%- endfor %}
        {%- endif %}
        {{- '<|im_end|>\n' }}
    {%- elif message.role == "tool" %}
        {%- if loop.previtem and loop.previtem.role != "tool" %}
            {{- '<|im_start|>user' }}
        {%- endif %}
        {{- '\n<tool_response>\n' }}
        {{- content }}
        {{- '\n</tool_response>' }}
        {%- if not loop.last and loop.nextitem.role != "tool" %}
            {{- '<|im_end|>\n' }}
        {%- elif loop.last %}
            {{- '<|im_end|>\n' }}
        {%- endif %}
    {%- else %}
        {{- raise_exception('Unexpected message role.') }}
    {%- endif %}
{%- endfor %}
{%- if add_generation_prompt %}
    {{- '<|im_start|>assistant\n' }}
    {%- if enable_thinking is defined and enable_thinking is false %}
        {{- '<think>\n\n</think>\n\n' }}
    {%- else %}
        {{- '<think>\n' }}
    {%- endif %}
{%- endif %}

Usage:
xxxx  --chat-template-file I:\aidraw\lm_models\codex.jinja --jinja

Core modifications：
{%- if message.role == "system" %}
        {%- if not loop.first %}
            {{- raise_exception('System message must be at the beginning.') }}
        {%- endif %}

Modified to

{%- if message.role == "system" %}
        {%- if not loop.first %}
            {{- '<|im_start|>user\n[System Instruction]\n' + content + '<|im_end|>' + '\n' }}
        {%- endif %}

[chenxqiyu]

Please reopen. This is a major incompatibility with OpenAI’s api.

After extracting templates and modifying them, the injection was fixed

[JamesFieldist]

Sorry for the dumb question, but is this a bug in llama.cpp or the unsloth model? If it's unsloth, I'd like to report it to them.

[rchildre3]

Not stale, I am hitting this with:

llama.cpp

• Version: b9444
• Build:

  cmake \
    -G Ninja \
    -B build \
    -DCMAKE_BUILD_TYPE=RelWithDebInfo \
    -DCMAKE_CXX_FLAGS="-march=native" \
    -DCMAKE_EXPORT_COMPILE_COMMANDS=1 \
    -DCMAKE_INTERPROCEDURAL_OPTIMIZATION=1 \
    -DGGML_HIP_ROCWMMA_FATTN=ON \
    -DGGML_HIP=ON \
    -DGGML_NATIVE=1 \
    -DGPU_TARGETS="gfx1101"
  

• Run: HIP_VISIBLE_DEVICES=0 ./build/bin/llama-server -hf unsloth/Qwen3.5-9B-GGUF -ngl 25 --threads 10

codex

• Version: codex-cli 0.135.0-alpha.1
• Config:

  model = "unsloth/Qwen3.5-9B-GGUF"
  model_provider = "llamacpp"
  
  [model_providers.llamacpp]
  name = "Local llama.cpp Server"
  base_url = "http://localhost:8080/v1"
  

Server error:

... W srv  server_chat_: unsupported Responses tool type 'custom' skipped
... W srv  server_chat_: unsupported Responses tool type 'tool_search' skipped
... W srv  server_chat_: unsupported Responses tool type 'namespace' skipped
... W srv  server_chat_: unsupported Responses tool type 'web_search' skipped
... W srv    operator(): got exception: {"error":{"code":400,"message":"Unable to generate parser for this template. Automatic parser generation failed: \n------------\nWhile executing CallExpression at line 85, column 32 in source:\n...first %}↵            {{- raise_exception('System message must be at the beginnin...\n                                           ^\nError: Jinja Exception: System message must be at the beginning.","type":"invalid_request_error"}}
... W srv    operator(): got exception: {"error":{"code":400,"message":"Unable to generate parser for this template. Automatic parser generation failed: \n------------\nWhile executing CallExpression at line 85, column 32 in source:\n...first %}↵            {{- raise_exception('System message must be at the beginnin...\n                                           ^\nError: Jinja Exception: System message must be at the beginning.","type":"invalid_request_error"}}

Edit:

This updated chat template seemed to have fixed Codex talking to the Qwen 3.5 model:

https://huggingface.co/froggeric/Qwen-Fixed-Chat-Templates

--jinja --chat-template-file chat_template.jinja