# Tool Calling

**URL:** https://discuss.vllm.ai/c/features/tool-calling/22.md

[Latest](https://discuss.vllm.ai/latest.md) · [Categories](https://discuss.vllm.ai/categories.md)

---

## [About the Tool Calling category](https://discuss.vllm.ai/t/about-the-tool-calling-category/90)

<div class="topic-metadata">

**Author:** [@DarkLight1337](https://discuss.vllm.ai/u/DarkLight1337)\
**Replies:** 0

</div>

---

## [DeepSeek-V3 tool\_choice="auto", not working but tool\_choice="required" is working](https://discuss.vllm.ai/t/deepseek-v3-tool-choice-auto-not-working-but-tool-choice-required-is-working/1006)

<div class="topic-metadata">

**Author:** [@Vipin](https://discuss.vllm.ai/u/Vipin)\
**Replies:** 4\
**Last updated:** [October 13, 2025, 4:11am UTC](https://discuss.vllm.ai/t/deepseek-v3-tool-choice-auto-not-working-but-tool-choice-required-is-working/1006 "2025-10-13T04:11:57Z")

</div>

DeepSeek-V3 tool\_choice=“auto”, not working but tool\_choice=“required” is working args: - --model - deepseek-ai/DeepSeek-V3 - --port - “8080” - --tensor-parallel-size - “8” - --served-model-name - deepseek\_v3 …

---

## [Error trying to handle streaming tool call](https://discuss.vllm.ai/t/error-trying-to-handle-streaming-tool-call/1104)

<div class="topic-metadata">

**Author:** [@XiaoDouGeGe](https://discuss.vllm.ai/u/XiaoDouGeGe)\
**Replies:** 3\
**Last updated:** [July 17, 2025, 7:49am UTC](https://discuss.vllm.ai/t/error-trying-to-handle-streaming-tool-call/1104 "2025-07-17T07:49:14Z")

</div>

vllm启动Qwen2.5-72B模型服务，能指定 --tool-call-parser hermes 参数配置吗？ 我的启动命令如下： export VLLM\_ATTENTION\_BACKEND=FLASHINFER python3 -m vllm.entrypoints.openai.api\_server --host 0.0.0.0 \\ --port 8000 \\ --model /model…

---

## [Tool calling using Offline Inference?](https://discuss.vllm.ai/t/tool-calling-using-offline-inference/452)

<div class="topic-metadata">

**Author:** [@bkuster](https://discuss.vllm.ai/u/bkuster)\
**Replies:** 1\
**Last updated:** [April 14, 2025, 3:40pm UTC](https://discuss.vllm.ai/t/tool-calling-using-offline-inference/452 "2025-04-14T15:40:00Z")

</div>

Hello, is it possible to use Tool Calling using the offline inference paradigm, e.g. using the LLM class? I noticed guided decoding supports JSON schema but in case of multiple tools, each obviously has a separate JSON…

---

## [Multiple tools with Mistral Large 2411](https://discuss.vllm.ai/t/multiple-tools-with-mistral-large-2411/217)

<div class="topic-metadata">

**Author:** [@DystopianJunkyardKid](https://discuss.vllm.ai/u/DystopianJunkyardKid)\
**Replies:** 4\
**Last updated:** [March 26, 2025, 2:34pm UTC](https://discuss.vllm.ai/t/multiple-tools-with-mistral-large-2411/217 "2025-03-26T14:34:28Z")

</div>

Please release an example of tools calling in a multiple tools environment. Currently ‘tools’ section only allowed single tool implementation, vllm server will omit ‘“POST /v1/chat/completions HTTP/1.1” 400 Bad Request’…