Spring AI when use deepseek-r1 mode in mcp call return the error of 400 Bad Request

Bug description when use deepseek-r1 mode in toolObjects call or mcp call both return the error of 400 Bad Request

Environment spring ai version: 1.0.0 deepseek-r1:8b deeployed by ollama in localhost

error info o.s.ai.chat.model.MessageAggregator : Aggregation Error org.springframework.web.reactive.function.client.WebClientResponseException$BadRequest: 400 Bad Request from POST http://localhost:11434/v1/chat/completions at org.springframework.web.reactive.function.client.WebClientResponseException.create(WebClientResponseException.java:321) ~[spring-webflux-6.2.6.jar:6.2.6] Suppressed: reactor.core.publisher.FluxOnAssembly$OnAssemblyException: Error has been observed at the following site(s): *__checkpoint ⇢ 400 BAD_REQUEST from POST http://localhost:11434/v1/chat/completions [DefaultWebClient] Original Stack Trace: at org.springframework.web.reactive.function.client.WebClientResponseException.create(WebClientResponseException.java:321) ~[spring-webflux-6.2.6.jar:6.2.6] at org.springframework.web.reactive.function.client.DefaultClientResponse.lambda$createException$1(DefaultClientResponse.java:214) ~[spring-webflux-6.2.6.jar:6.2.6] at reactor.core.publisher.FluxMap$MapSubscriber.onNext(FluxMap.java:106) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxOnErrorReturn$ReturnSubscriber.onNext(FluxOnErrorReturn.java:162) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxDefaultIfEmpty$DefaultIfEmptySubscriber.onNext(FluxDefaultIfEmpty.java:122) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxMapFuseable$MapFuseableSubscriber.onNext(FluxMapFuseable.java:129) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxContextWrite$ContextWriteSubscriber.onNext(FluxContextWrite.java:107) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxMapFuseable$MapFuseableConditionalSubscriber.onNext(FluxMapFuseable.java:299) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxFilterFuseable$FilterFuseableConditionalSubscriber.onNext(FluxFilterFuseable.java:337) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.Operators$BaseFluxToMonoOperator.completePossiblyEmpty(Operators.java:2096) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.MonoCollect$CollectSubscriber.onComplete(MonoCollect.java:145) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxMap$MapSubscriber.onComplete(FluxMap.java:144) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxPeek$PeekSubscriber.onComplete(FluxPeek.java:260) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.core.publisher.FluxMap$MapSubscriber.onComplete(FluxMap.java:144) ~[reactor-core-3.7.5.jar:3.7.5] at reactor.netty.channel.FluxReceive.onInboundComplete(FluxReceive.java:413) ~[reactor-netty-core-1.2.5.jar:1.2.5] at reactor.netty.channel.ChannelOperations.onInboundComplete(ChannelOperations.java:455) ~[reactor-netty-core-1.2.5.jar:1.2.5] at reactor.netty.channel.ChannelOperations.terminate(ChannelOperations.java:509) ~[reactor-netty-core-1.2.5.jar:1.2.5] at reactor.netty.http.client.HttpClientOperations.onInboundNext(HttpClientOperations.java:821) ~[reactor-netty-http-1.2.5.jar:1.2.5] at reactor.netty.channel.ChannelOperationsHandler.channelRead(ChannelOperationsHandler.java:115) ~[reactor-netty-core-1.2.5.jar:1.2.5] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:444) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:420) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:412) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.handler.codec.MessageToMessageDecoder.channelRead(MessageToMessageDecoder.java:107) ~[netty-codec-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:444) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:420) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:412) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.CombinedChannelDuplexHandler$DelegatingChannelHandlerContext.fireChannelRead(CombinedChannelDuplexHandler.java:436) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.handler.codec.ByteToMessageDecoder.fireChannelRead(ByteToMessageDecoder.java:346) ~[netty-codec-4.1.119.Final.jar:4.1.119.Final] at io.netty.handler.codec.ByteToMessageDecoder.channelRead(ByteToMessageDecoder.java:318) ~[netty-codec-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.CombinedChannelDuplexHandler.channelRead(CombinedChannelDuplexHandler.java:251) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:442) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:420) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:412) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.DefaultChannelPipeline$HeadContext.channelRead(DefaultChannelPipeline.java:1357) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:440) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:420) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.DefaultChannelPipeline.fireChannelRead(DefaultChannelPipeline.java:868) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.nio.AbstractNioByteChannel$NioByteUnsafe.read(AbstractNioByteChannel.java:166) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:796) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:732) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:658) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:562) ~[netty-transport-4.1.119.Final.jar:4.1.119.Final] at io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:998) ~[netty-common-4.1.119.Final.jar:4.1.119.Final] at io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74) ~[netty-common-4.1.119.Final.jar:4.1.119.Final] at io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30) ~[netty-common-4.1.119.Final.jar:4.1.119.Final] at java.base/java.lang.Thread.run(Thread.java:840) ~[na:na]

Comment From: sunyuhan1998

Actually this has nothing to do with MCP, it's because in Ollama the deepseek model is not supported for tool calls.

If you try to send an HTTP request with tools directly to the REST interface of Ollama:

curl -X POST http://localhost:11434/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-r1:8b",
    "messages": [
      {
        "role": "user",
        "content": "What time is it now?？\n"
      }
    ],
    "stream": true,
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "getCurrentDateTime",
          "description": "Get the current date and time in the user'\''s timezone",
          "parameters": {
            "$schema": "https://json-schema.org/draft/2020-12/schema",
            "additionalProperties": false,
            "type": "object",
            "properties": {},
            "required": []
          }
        }
      }
    ],
    "options": {}
  }'

you will get the following error:

{"error":"registry.ollama.ai/library/deepseek-r1:8b does not support tools"}

You can find the models that Ollama supports for tool calls at https://ollama.com/search?c=tools , and you can see that so far the official deepseek model is not among them.

But actually, the description in deepseek's official api documentation: https://api-docs.deepseek.com/guides/reasoning_model#api-parameters, deepseek itself supports tool calls (but I'm not sure which size of its open source model can support tool calls)

Overall, I'm inclined to think it's an Ollama support issue.

Comment From: MochaSoft2023

Thank you very much for your reply, which has helped me solve my long-standing confusion. I will try other methods of model selection and deployment. Thank you again

Comment From: maruixiang

In earlier version for ollama and deepseek-r1, the error is "does not support tools". But in ollama 0.9 and DeepSeek-R1-0528-Qwen3-8B , the error is "400 Bad Request... Aggregation Error" . They are not the same error.

Comment From: sunyuhan1998

In earlier version for ollama and deepseek-r1, the error is "does not support tools". But in ollama 0.9 and DeepSeek-R1-0528-Qwen3-8B , the error is "400 Bad Request... Aggregation Error" . They are not the same error.

Not really, if you try to capture the body of the exception content returned by OllamaApi's webClient, you'll see that it's actually still the "does not support tools" error.