[Feature] Gigachat 3.5 support (#29189)

Co-authored-by: Stanislav Petrov <stapetrov@sberbank.ru>
Co-authored-by: Viacheslav Barinov <vvadbarinov@sberbank.ru>
Co-authored-by: Viacheslav <viacheslav.teh@gmail.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>

docs/docs/advanced_features/server_arguments.mdx

@@ -1154,13 +1154,13 @@ Combining `--enable-response-store` with `--disaggregation-mode=prefill` or `dec
       <td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}>`--reasoning-parser`</td>
       <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}>Specify the parser for reasoning models. Use `auto` to detect the parser from the model's chat template.</td>
       <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}>`None`</td>
-      <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>auto</code>, <code>apertus2509</code>, <code>deepseek-r1</code>, <code>deepseek-v3</code>, <code>deepseek-v4</code>, <code>dots</code>, <code>glm45</code>, <code>ling3</code>, <code>hunyuan</code>, <code>gpt-oss</code>, <code>k2_horizon</code>, <code>kimi</code>, <code>kimi_k2</code>, <code>kimi_k3</code>, <code>mimo</code>, <code>muse</code>, <code>poolside_v1</code>, <code>qwen3</code>, <code>qwen3-thinking</code>, <code>minimax</code>, <code>minimax-append-think</code>, <code>minimax-m3</code>, <code>step3</code>, <code>step3p5</code>, <code>mistral</code>, <code>nemotron_3</code>, <code>interns1</code>, <code>gemma4</code>, <code>inkling</code>, <code>cohere_command4</code></td>
+      <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>auto</code>, <code>apertus2509</code>, <code>deepseek-r1</code>, <code>deepseek-v3</code>, <code>deepseek-v4</code>, <code>dots</code>, <code>glm45</code>, <code>ling3</code>, <code>hunyuan</code>, <code>gpt-oss</code>, <code>k2_horizon</code>, <code>kimi</code>, <code>kimi_k2</code>, <code>kimi_k3</code>, <code>mimo</code>, <code>muse</code>, <code>poolside_v1</code>, <code>qwen3</code>, <code>qwen3-thinking</code>, <code>minimax</code>, <code>minimax-append-think</code>, <code>minimax-m3</code>, <code>step3</code>, <code>step3p5</code>, <code>mistral</code>, <code>nemotron_3</code>, <code>interns1</code>, <code>gemma4</code>, <code>gigachat35</code>, <code>inkling</code>, <code>cohere_command4</code></td>
     </tr>
         <tr>
       <td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}>`--tool-call-parser`</td>
       <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}>Specify the parser for handling tool-call interactions. Use `auto` to detect the parser from the model's chat template.</td>
       <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}>`None`</td>
-      <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>auto</code>, <code>apertus2509</code>, <code>cohere_command4</code>, <code>deepseekv3</code>, <code>deepseekv31</code>, <code>deepseekv32</code>, <code>deepseekv4</code>, <code>dots</code>, <code>glm</code>, <code>glm45</code>, <code>glm47</code>, <code>gpt-oss</code>, <code>k2_horizon</code>, <code>kimi_k2</code>, <code>kimi_k3</code>, <code>lfm2</code>, <code>ling3</code>, <code>llama3</code>, <code>mimo</code>, <code>minicpm5</code>, <code>mistral</code>, <code>muse</code>, <code>poolside_v1</code>, <code>pythonic</code>, <code>qwen</code>, <code>qwen25</code>, <code>qwen3_coder</code>, <code>spark25</code>, <code>step3</code>, <code>step3p5</code>, <code>minimax-m2</code>, <code>minimax-m3</code>, <code>trinity</code>, <code>interns1</code>, <code>hermes</code>, <code>hunyuan</code>, <code>gigachat3</code>, <code>gemma4</code>, <code>inkling</code></td>
+      <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>auto</code>, <code>apertus2509</code>, <code>cohere_command4</code>, <code>deepseekv3</code>, <code>deepseekv31</code>, <code>deepseekv32</code>, <code>deepseekv4</code>, <code>dots</code>, <code>glm</code>, <code>glm45</code>, <code>glm47</code>, <code>gpt-oss</code>, <code>k2_horizon</code>, <code>kimi_k2</code>, <code>kimi_k3</code>, <code>lfm2</code>, <code>ling3</code>, <code>llama3</code>, <code>mimo</code>, <code>minicpm5</code>, <code>mistral</code>, <code>muse</code>, <code>poolside_v1</code>, <code>pythonic</code>, <code>qwen</code>, <code>qwen25</code>, <code>qwen3_coder</code>, <code>spark25</code>, <code>step3</code>, <code>step3p5</code>, <code>minimax-m2</code>, <code>minimax-m3</code>, <code>trinity</code>, <code>interns1</code>, <code>hermes</code>, <code>hunyuan</code>, <code>gigachat3</code>, <code>gigachat35</code>, <code>gemma4</code>, <code>inkling</code></td>
     </tr>
         <tr>
       <td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}>`--tool-server`</td>

docs/docs/supported-models/generative_models.mdx

@@ -308,5 +308,10 @@ in the GitHub search bar.
       <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>JetBrains/Mellum2-12B-A2.5B-Base</code>, <code>JetBrains/Mellum2-12B-A2.5B-Thinking</code></td>
       <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}>JetBrains' Qwen3-MoE-based code generation model with interleaved sliding-window/full attention, per-layer-type RoPE, and per-layer dense/sparse MLP routing.</td>
     </tr>
+<tr>
+      <td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}><strong>GigaChat 3.5</strong></td>
+      <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}><code>ai-sage/GigaChat3.5-432B-A28B</code></td>
+      <td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}>GigaChat's hybrid MoE model: per-layer mix of DeepSeek-style MLA full attention and Qwen3 Gated-Delta-Net (GDN) linear attention, DeepSeek-style routed + shared experts, and multiple multi-token-prediction (MTP) heads for speculative decoding. Emits tool calls in GCML format.</td>
+    </tr>
 </tbody>
 </table>