feat(web): add Xiaomi MiMo native search (rescue of #5687)

CHANGELOG.md

@@ -9,6 +9,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+- Add provider-native web search for documented Xiaomi MiMo 2.5 Pro and 2.5
+  chat routes while keeping neighboring models and custom gateways fail-closed.
 - Add structured provider-native web search for exact Z.AI global and Zhipu
   China general API routes, using each site's documented search engine value.
   Existing `open.bigmodel.cn` configurations now share Z.AI's official model

crates/config/src/route/capabilities.rs

@@ -52,6 +52,7 @@ impl CapabilityState {
 /// - OpenAI Responses web search: <https://developers.openai.com/api/docs/guides/tools-web-search>
 /// - Anthropic web search tool: <https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool>
 /// - xAI web search tool: <https://docs.x.ai/developers/tools/web-search>
+/// - Xiaomi MiMo web search: <https://mimo.mi.com/docs/en-US/usage-guide/tool-calling/web-search>
 /// - Z.AI Web Search API: <https://docs.z.ai/api-reference/tools/web-search>
 /// - Zhipu Web Search API: <https://docs.bigmodel.cn/api-reference/工具-api/网络搜索>
 /// - Alibaba Model Studio Token Plan Harness tools: <https://help.aliyun.com/en/model-studio/token-plan-harness-tool>
@@ -80,6 +81,7 @@ pub(crate) fn documented_server_side_web_search(
                 | "claude-sonnet-4-6"
         ),
         "xai" => matches!(wire_model_id.as_str(), "grok-4.6" | "grok-4.5"),
+        "xiaomi-mimo" => matches!(wire_model_id.as_str(), "mimo-v2.5-pro" | "mimo-v2.5"),
         "zai" => matches!(
             wire_model_id.as_str(),
             "glm-5.3" | "glm-5.3-flash" | "glm-5.2" | "glm-5.1" | "glm-5-turbo"
@@ -198,6 +200,10 @@ mod tests {
             documented_server_side_web_search("anthropic", "claude-sonnet-4-6"),
             CapabilityState::Supported
         );
+        assert_eq!(
+            documented_server_side_web_search("xiaomi-mimo", "mimo-v2.5-pro"),
+            CapabilityState::Supported
+        );
         assert_eq!(
             documented_server_side_web_search("zai", "GLM-5.3"),
             CapabilityState::Supported
@@ -219,6 +225,7 @@ mod tests {
             ("xai", "grok-4.6-latest"),
             ("xai", "grok-4.5-fast"),
             ("anthropic", "claude-haiku-4-5"),
+            ("xiaomi-mimo", "mimo-v2.5-pro-ultraspeed"),
             ("zai", "glm-5.3-preview"),
             ("modelstudio-coding-plan", "qwen3.8-max"),
             ("modelstudio-token-plan", "qwen3.8-max-preview"),

crates/config/src/route/tests.rs

@@ -1433,6 +1433,47 @@ fn provider_native_web_search_requires_exact_direct_endpoint_offering() {
     );
 }
 
+#[test]
+fn mimo_native_search_is_exact_to_documented_chat_models() {
+    use crate::route::CapabilityState;
+
+    let resolver = RouteResolver::new();
+    for model in ["mimo-v2.5-pro", "mimo-v2.5"] {
+        let direct = resolver
+            .resolve(&req(Some(ProviderKind::XiaomiMimo), Some(model)))
+            .expect("official MiMo chat route resolves");
+        assert_eq!(
+            direct.capabilities().server_side_web_search,
+            CapabilityState::Supported
+        );
+    }
+
+    let neighboring = resolver
+        .resolve(&req(
+            Some(ProviderKind::XiaomiMimo),
+            Some("mimo-v2.5-pro-ultraspeed"),
+        ))
+        .expect("neighboring MiMo model resolves");
+    assert_eq!(
+        neighboring.capabilities().server_side_web_search,
+        CapabilityState::Unknown
+    );
+
+    let custom = resolver
+        .resolve(&RouteRequest {
+            explicit_provider: Some(ProviderKind::XiaomiMimo),
+            model_selector: Some(LogicalModelRef::from("mimo-v2.5-pro")),
+            saved_provider_model: None,
+            base_url_override: Some("https://compatible.example.test/v1".to_string()),
+            limit_overrides: Vec::new(),
+        })
+        .expect("custom MiMo-compatible route resolves");
+    assert_eq!(
+        custom.capabilities().server_side_web_search,
+        CapabilityState::Unknown
+    );
+}
+
 #[test]
 fn zai_native_search_requires_exact_general_api_product() {
     use crate::route::CapabilityState;

crates/tui/CHANGELOG.md

@@ -9,6 +9,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+- Add provider-native web search for documented Xiaomi MiMo 2.5 Pro and 2.5
+  chat routes while keeping neighboring models and custom gateways fail-closed.
 - Add structured provider-native web search for exact Z.AI global and Zhipu
   China general API routes, using each site's documented search engine value.
   Existing `open.bigmodel.cn` configurations now share Z.AI's official model

crates/tui/src/client/provider_native_search.rs

@@ -50,6 +50,7 @@ impl ProviderNativeSearchClient {
             ApiProvider::Openai
                 | ApiProvider::Anthropic
                 | ApiProvider::Xai
+                | ApiProvider::XiaomiMimo
                 | ApiProvider::Zai
                 | ApiProvider::ModelstudioTokenPlan
                 | ApiProvider::Deepseek
@@ -141,6 +142,7 @@ impl ProviderNativeSearchClient {
                     2_048_u32.min(route_cap),
                 )
             }
+            ApiProvider::XiaomiMimo => build_mimo_search_body(&self.inner.default_model, request),
             ApiProvider::Zai => zai::build_body(request, &self.inner.base_url)?,
             _ => bail!("active provider has no native web-search adapter"),
         };
@@ -152,6 +154,7 @@ impl ProviderNativeSearchClient {
                 responses_api_url(&self.inner.base_url, self.inner.api_provider)
             }
             ApiProvider::Anthropic => anthropic_messages_url(&self.inner.base_url),
+            ApiProvider::XiaomiMimo => api_url(&self.inner.base_url, "chat/completions"),
             ApiProvider::Zai => api_url(&self.inner.base_url, "web_search"),
             _ => unreachable!("provider checked above"),
         };
@@ -179,6 +182,7 @@ impl ProviderNativeSearchClient {
             | ApiProvider::Deepseek
             | ApiProvider::DeepseekCN => parse_responses_search(&payload),
             ApiProvider::Anthropic => parse_anthropic_search(&payload),
+            ApiProvider::XiaomiMimo => parse_mimo_search(&payload),
             ApiProvider::Zai => zai::parse(&payload),
             _ => unreachable!("provider checked above"),
         };
@@ -262,6 +266,23 @@ fn build_anthropic_search_body(
     })
 }
 
+fn build_mimo_search_body(model: &str, request: &ProviderNativeSearchRequest) -> Value {
+    json!({
+        "model": model,
+        "messages": [{ "role": "user", "content": search_prompt(request) }],
+        "tools": [{
+            "type": "web_search",
+            "max_keyword": 1,
+            "force_search": true,
+            "limit": request.max_results,
+        }],
+        "tool_choice": "auto",
+        "max_completion_tokens": 2_048,
+        "stream": false,
+        "thinking": { "type": "disabled" },
+    })
+}
+
 fn anthropic_messages_url(base_url: &str) -> String {
     let base = base_url.trim_end_matches('/');
     if base.ends_with("/v1") {
@@ -388,6 +409,48 @@ fn parse_anthropic_search(payload: &Value) -> ProviderNativeSearchResponse {
     }
 }
 
+fn parse_mimo_search(payload: &Value) -> ProviderNativeSearchResponse {
+    let message = payload.pointer("/choices/0/message");
+    let answer = message
+        .and_then(|value| value.get("content"))
+        .and_then(Value::as_str)
+        .map(str::trim)
+        .filter(|text| !text.is_empty())
+        .map(str::to_string);
+    let mut citations = Vec::new();
+    if let Some(annotations) = message
+        .and_then(|value| value.get("annotations"))
+        .and_then(Value::as_array)
+        .or_else(|| payload.get("annotations").and_then(Value::as_array))
+    {
+        for annotation in annotations {
+            let Some(url) = annotation.get("url").and_then(Value::as_str) else {
+                continue;
+            };
+            let title = annotation
+                .get("title")
+                .and_then(Value::as_str)
+                .map(str::to_string);
+            let snippet = annotation
+                .get("summary")
+                .and_then(Value::as_str)
+                .map(str::to_string);
+            let published = annotation
+                .get("publish_time")
+                .and_then(Value::as_str)
+                .map(str::to_string);
+            push_citation(
+                &mut citations,
+                citation_from_url(url, title, snippet, published),
+            );
+        }
+    }
+    ProviderNativeSearchResponse {
+        answer: bounded_answer(answer.into_iter().collect()),
+        citations,
+    }
+}
+
 fn citation_from_value(
     value: &Value,
     snippet: Option<String>,
@@ -596,6 +659,16 @@ mod tests {
         assert_eq!(tiny_route["max_tokens"], 128);
     }
 
+    #[test]
+    fn mimo_payload_forces_bounded_web_search_plugin() {
+        let body = build_mimo_search_body("mimo-v2.5-pro", &request());
+        assert_eq!(body["tools"][0]["type"], "web_search");
+        assert_eq!(body["tools"][0]["force_search"], true);
+        assert_eq!(body["tools"][0]["limit"], 3);
+        assert_eq!(body["max_completion_tokens"], 2_048);
+        assert_eq!(body["thinking"]["type"], "disabled");
+    }
+
     #[test]
     fn responses_parser_separates_answer_and_deduplicated_citations() {
         let payload = json!({
@@ -706,6 +779,28 @@ mod tests {
         );
     }
 
+    #[test]
+    fn mimo_parser_keeps_non_streaming_annotations() {
+        let parsed = parse_mimo_search(&json!({
+            "choices": [{
+                "message": {
+                    "content": "Grounded answer.",
+                    "annotations": [{
+                        "type": "url_citation",
+                        "url": "https://example.com/weather",
+                        "title": "Weather",
+                        "summary": "Forecast",
+                        "publish_time": "2026-08-28"
+                    }]
+                }
+            }]
+        }));
+        assert_eq!(parsed.answer.as_deref(), Some("Grounded answer."));
+        assert_eq!(parsed.citations.len(), 1);
+        assert_eq!(parsed.citations[0].snippet.as_deref(), Some("Forecast"));
+        assert_eq!(parsed.citations[0].published.as_deref(), Some("2026-08-28"));
+    }
+
     #[test]
     fn non_http_citations_are_rejected() {
         let payload = json!({ "citations": ["javascript:alert(1)"] });

docs/PROVIDERS.md

@@ -572,7 +572,7 @@ overlay and lets DSH resolve its own keys.
 | `volcengine` | `[providers.volcengine]` | `VOLCENGINE_API_KEY`, `VOLCENGINE_ARK_API_KEY`, `ARK_API_KEY` | `VOLCENGINE_BASE_URL`, `VOLCENGINE_ARK_BASE_URL`, `ARK_BASE_URL`; default `https://ark.cn-beijing.volces.com/api/coding/v3` | `DeepSeek-V4-Pro`, `DeepSeek-V4-Flash` | Volcengine/Volcano Engine Ark OpenAI-compatible coding endpoint. `VOLCENGINE_MODEL` and `VOLCENGINE_ARK_MODEL` are accepted. |
 | `openrouter` | `[providers.openrouter]` | `OPENROUTER_API_KEY` | `OPENROUTER_BASE_URL`; default `https://openrouter.ai/api/v1` | `deepseek/deepseek-v4-pro`, `deepseek/deepseek-v4-flash`; recent large IDs include `arcee-ai/trinity-large-thinking`, `minimax/minimax-m3`, `xiaomi/mimo-v2.5-pro`, `qwen/qwen3.6-flash`, `qwen/qwen3.6-35b-a3b`, `qwen/qwen3.6-max-preview`, `qwen/qwen3.6-27b`, `qwen/qwen3.6-plus`, `google/gemma-4-31b-it`, `z-ai/glm-5.1`, `z-ai/glm-5.2`, `moonshotai/kimi-k2.7-code`, `moonshotai/kimi-k2.6` | Additive open-model routing layer. It does not replace DeepSeek; it lets users route supported model IDs through OpenRouter when they choose it. |
 | `orcarouter` | `[providers.orcarouter]` | `ORCAROUTER_API_KEY` | `ORCAROUTER_BASE_URL`; default `https://api.orcarouter.ai/v1` | `deepseek/deepseek-v4-pro`, `deepseek/deepseek-v4-flash`; router alias `orcarouter/auto`; recent large IDs mirror the OpenRouter namespaced catalog | [OrcaRouter](https://www.orcarouter.ai) OpenAI-compatible aggregation gateway. Shares the namespaced `vendor/model` wire-model format and DeepSeek model set with OpenRouter, so the OpenRouter base-URL and model-normalization rules apply. `ORCAROUTER_MODEL` is accepted. Provider aliases: `orcarouter`, `orca_router`, `orca`. |
-| `xiaomi-mimo` | `[providers.xiaomi_mimo]` | `XIAOMI_MIMO_TOKEN_PLAN_API_KEY`, `MIMO_TOKEN_PLAN_API_KEY`, `XIAOMI_MIMO_API_KEY`, `XIAOMI_API_KEY`, `MIMO_API_KEY` | `XIAOMI_MIMO_BASE_URL`, `MIMO_BASE_URL`, `XIAOMI_MIMO_MODE`, `MIMO_MODE`; default `https://token-plan-sgp.xiaomimimo.com/v1` | Chat: `mimo-v2.5-pro`, `mimo-v2.5-pro-ultraspeed`, `mimo-v2.5`; speech/TTS: `mimo-v2.5-tts`, `mimo-v2.5-tts-voicedesign`, `mimo-v2.5-tts-voiceclone`, `mimo-v2-tts` | Xiaomi MiMo OpenAI-compatible chat completions route. Token Plan keys (`tp-...`) use `api-key` auth and the token-plan endpoint by default; pay-as-you-go mode uses standard API keys (`sk-...`) and `https://api.xiaomimimo.com/v1`. It sends `max_completion_tokens` and uses MiMo's `thinking` field for reasoning control. Token Plan cost/usage is credit/quota based; Codewhale shows it as unknown until Xiaomi exposes a reliable balance API. `codewhale speech` / `tts` uses the TTS models. |
+| `xiaomi-mimo` | `[providers.xiaomi_mimo]` | `XIAOMI_MIMO_TOKEN_PLAN_API_KEY`, `MIMO_TOKEN_PLAN_API_KEY`, `XIAOMI_MIMO_API_KEY`, `XIAOMI_API_KEY`, `MIMO_API_KEY` | `XIAOMI_MIMO_BASE_URL`, `MIMO_BASE_URL`, `XIAOMI_MIMO_MODE`, `MIMO_MODE`; default `https://token-plan-sgp.xiaomimimo.com/v1` | Chat: `mimo-v2.5-pro`, `mimo-v2.5-pro-ultraspeed`, `mimo-v2.5`; speech/TTS: `mimo-v2.5-tts`, `mimo-v2.5-tts-voicedesign`, `mimo-v2.5-tts-voiceclone`, `mimo-v2-tts` | Xiaomi MiMo OpenAI-compatible chat completions route. `mimo-v2.5-pro` and `mimo-v2.5` can use the documented provider-native web-search plugin; ultraspeed, speech/TTS, and custom-compatible routes do not inherit it. Token Plan keys (`tp-...`) use `api-key` auth and the token-plan endpoint by default; pay-as-you-go mode uses standard API keys (`sk-...`) and `https://api.xiaomimimo.com/v1`. It sends `max_completion_tokens` and uses MiMo's `thinking` field for reasoning control. Token Plan cost/usage is credit/quota based; Codewhale shows it as unknown until Xiaomi exposes a reliable balance API. `codewhale speech` / `tts` uses the TTS models. |
 | `novita` | `[providers.novita]` | `NOVITA_API_KEY` | `NOVITA_BASE_URL`; default `https://api.novita.ai/openai/v1` | `deepseek/deepseek-v4-pro`, `deepseek/deepseek-v4-flash` | OpenAI-compatible hosted route for DeepSeek model IDs. Use config or `CODEWHALE_MODEL` / `DEEPSEEK_MODEL` for model overrides. |
 | `fireworks` | `[providers.fireworks]` | `FIREWORKS_API_KEY` | `FIREWORKS_BASE_URL`; default `https://api.fireworks.ai/inference/v1` | `accounts/fireworks/models/deepseek-v4-pro` | OpenAI-compatible hosted route. Use config or `CODEWHALE_MODEL` / `DEEPSEEK_MODEL` for model overrides. |
 | `siliconflow` | `[providers.siliconflow]` | `SILICONFLOW_API_KEY` | `SILICONFLOW_BASE_URL`; default `https://api.siliconflow.com/v1` | `deepseek-ai/DeepSeek-V4-Pro`, `deepseek-ai/DeepSeek-V4-Flash` | OpenAI-compatible hosted route. Official docs use the `.com` endpoint. `SILICONFLOW_MODEL` is accepted. Reasoning aliases `deepseek-reasoner` and `deepseek-r1` map to Pro; `deepseek-chat` and `deepseek-v3` map to Flash. |