add trinity-large-preview & openrouter/stepfun/step-3.5-flash

index.html

@@ -1246,8 +1246,44 @@
                         <td class="num mono" data-label="Всего токенов">157,197</td>
                     </tr>
                     
-                    <tr data-model="openai/tiiuae/Falcon-H1-34B-Instruct">
+                    <tr data-model="openrouter/arcee-ai/trinity-large-preview:free">
                         <td class="rank mono sticky-0" data-label="#">#33</td>
+                        <td class="model-name sticky-1" data-label="Модель">arcee-ai/Trinity-Large-Preview 0.6 (Arcee-AI API)</td>
+                        <td class="num mono" data-label="Критичные/1000">
+                            
+                            0.22 ± 0.04
+                            
+                        </td>
+                        <td class="num mono" data-label="Обычные/1000">
+                            
+                            0.73 ± 0.24
+                            
+                        </td>
+                        <td class="num mono" data-label="Доп./1000">
+                            
+                            0.42 ± 0.01
+                            
+                        </td>
+                        <td data-label="Нормировано ошибок">
+                            <div class="score-cell">
+                                <div class="progress-bar">
+                                    
+                                    
+                                    
+                                    <div class="progress-fill" style="width: 65.25000000000001%"></div>
+                                </div>
+                                <span class="score-value">
+                                    
+                                    1.39 ± 0.21
+                                    
+                                </span>
+                            </div>
+                        </td>
+                        <td class="num mono" data-label="Всего токенов">61,298</td>
+                    </tr>
+                    
+                    <tr data-model="openai/tiiuae/Falcon-H1-34B-Instruct">
+                        <td class="rank mono sticky-0" data-label="#">#34</td>
                         <td class="model-name sticky-1" data-label="Модель">tiiuae/Falcon-H1-34B-Instruct (vllm)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1283,7 +1319,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/z-ai/glm-4.7-flash">
-                        <td class="rank mono sticky-0" data-label="#">#34</td>
+                        <td class="rank mono sticky-0" data-label="#">#35</td>
                         <td class="model-name sticky-1" data-label="Модель">zai-org/GLM-4.7-Flash (z.ai API, reasoning, temp=0.6)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1319,7 +1355,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/qwen/qwen3-235b-a22b-2507">
-                        <td class="rank mono sticky-0" data-label="#">#35</td>
+                        <td class="rank mono sticky-0" data-label="#">#36</td>
                         <td class="model-name sticky-1" data-label="Модель">Qwen3-235B-A22B-2507-Instruct (Alibaba API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1355,7 +1391,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/qwen/qwen3-vl-8b-instruct">
-                        <td class="rank mono sticky-0" data-label="#">#36</td>
+                        <td class="rank mono sticky-0" data-label="#">#37</td>
                         <td class="model-name sticky-1" data-label="Модель">Qwen3-VL-8B-Instruct (Alibaba API, presence_penalty=2)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1391,7 +1427,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/moonshotai/kimi-k2.5">
-                        <td class="rank mono sticky-0" data-label="#">#37</td>
+                        <td class="rank mono sticky-0" data-label="#">#38</td>
                         <td class="model-name sticky-1" data-label="Модель">moonshotai/Kimi-K2.5 (reasoning, temp=1, Moonshot AI API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1427,7 +1463,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/moonshotai/kimi-k2-0905">
-                        <td class="rank mono sticky-0" data-label="#">#38</td>
+                        <td class="rank mono sticky-0" data-label="#">#39</td>
                         <td class="model-name sticky-1" data-label="Модель">moonshotai/Kimi-K2-Instruct-0905 (Novita API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1463,7 +1499,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/z-ai/glm-4.6">
-                        <td class="rank mono sticky-0" data-label="#">#39</td>
+                        <td class="rank mono sticky-0" data-label="#">#40</td>
                         <td class="model-name sticky-1" data-label="Модель">GLM-4.6 (Z.ai API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1498,8 +1534,44 @@
                         <td class="num mono" data-label="Всего токенов">151,775</td>
                     </tr>
                     
+                    <tr data-model="openrouter/arcee-ai/trinity-large-preview:free">
+                        <td class="rank mono sticky-0" data-label="#">#41</td>
+                        <td class="model-name sticky-1" data-label="Модель">arcee-ai/Trinity-Large-Preview (Arcee-AI API)</td>
+                        <td class="num mono" data-label="Критичные/1000">
+                            
+                            0.23 ± 0.02
+                            
+                        </td>
+                        <td class="num mono" data-label="Обычные/1000">
+                            
+                            1.22 ± 0.59
+                            
+                        </td>
+                        <td class="num mono" data-label="Доп./1000">
+                            
+                            0.26 ± 0.05
+                            
+                        </td>
+                        <td data-label="Нормировано ошибок">
+                            <div class="score-cell">
+                                <div class="progress-bar">
+                                    
+                                    
+                                    
+                                    <div class="progress-fill" style="width: 54.75%"></div>
+                                </div>
+                                <span class="score-value">
+                                    
+                                    1.81 ± 0.55
+                                    
+                                </span>
+                            </div>
+                        </td>
+                        <td class="num mono" data-label="Всего токенов">80,598</td>
+                    </tr>
+                    
                     <tr data-model="openrouter/openai/gpt-5">
-                        <td class="rank mono sticky-0" data-label="#">#40</td>
+                        <td class="rank mono sticky-0" data-label="#">#42</td>
                         <td class="model-name sticky-1" data-label="Модель">GPT-5 (reasoning: minimal)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1535,7 +1607,7 @@
                     </tr>
                     
                     <tr data-model="openai/aquif-3.5-Max-42B-A3B.Q8_0.gguf">
-                        <td class="rank mono sticky-0" data-label="#">#41</td>
+                        <td class="rank mono sticky-0" data-label="#">#43</td>
                         <td class="model-name sticky-1" data-label="Модель">aquif-ai/aquif-3.5-Plus-30B-A3B (Q8_0 llama.cpp, without reasoning)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1571,7 +1643,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/openai/gpt-5">
-                        <td class="rank mono sticky-0" data-label="#">#42</td>
+                        <td class="rank mono sticky-0" data-label="#">#44</td>
                         <td class="model-name sticky-1" data-label="Модель">GPT-5 (reasoning: low)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1607,7 +1679,7 @@
                     </tr>
                     
                     <tr data-model="openai/nvidia/NVIDIA-Nemotron-Nano-12B-v2">
-                        <td class="rank mono sticky-0" data-label="#">#43</td>
+                        <td class="rank mono sticky-0" data-label="#">#45</td>
                         <td class="model-name sticky-1" data-label="Модель">nvidia/NVIDIA-Nemotron-Nano-12B-v2 (vllm, reasoning=false)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1642,8 +1714,44 @@
                         <td class="num mono" data-label="Всего токенов">86,045</td>
                     </tr>
                     
+                    <tr data-model="openrouter/stepfun/step-3.5-flash:free">
+                        <td class="rank mono sticky-0" data-label="#">#46</td>
+                        <td class="model-name sticky-1" data-label="Модель">Step-3.5-flash (StepFun API, fp8, temp=0.6, reasoning)</td>
+                        <td class="num mono" data-label="Критичные/1000">
+                            
+                            0.60 ± 0.05
+                            
+                        </td>
+                        <td class="num mono" data-label="Обычные/1000">
+                            
+                            0.85 ± 0.06
+                            
+                        </td>
+                        <td class="num mono" data-label="Доп./1000">
+                            
+                            0.27 ± 0.02
+                            
+                        </td>
+                        <td data-label="Нормировано ошибок">
+                            <div class="score-cell">
+                                <div class="progress-bar">
+                                    
+                                    
+                                    
+                                    <div class="progress-fill" style="width: 45.75%"></div>
+                                </div>
+                                <span class="score-value">
+                                    
+                                    2.17 ± 0.06
+                                    
+                                </span>
+                            </div>
+                        </td>
+                        <td class="num mono" data-label="Всего токенов">140,382</td>
+                    </tr>
+                    
                     <tr data-model="openrouter/openai/gpt-oss-120b">
-                        <td class="rank mono sticky-0" data-label="#">#44</td>
+                        <td class="rank mono sticky-0" data-label="#">#47</td>
                         <td class="model-name sticky-1" data-label="Модель">GPT-OSS-120B (Vertex AI API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1679,7 +1787,7 @@
                     </tr>
                     
                     <tr data-model="openai/gpt-4.1">
-                        <td class="rank mono sticky-0" data-label="#">#45</td>
+                        <td class="rank mono sticky-0" data-label="#">#48</td>
                         <td class="model-name sticky-1" data-label="Модель">GPT-4.1</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1715,7 +1823,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/nvidia/nemotron-3-nano-30b-a3b:free">
-                        <td class="rank mono sticky-0" data-label="#">#46</td>
+                        <td class="rank mono sticky-0" data-label="#">#49</td>
                         <td class="model-name sticky-1" data-label="Модель">Nemotron-3-Nano-30B-A3B (NVIDIA API, non-reasoning)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1751,7 +1859,7 @@
                     </tr>
                     
                     <tr data-model="mistral/ministral-3b-2512">
-                        <td class="rank mono sticky-0" data-label="#">#47</td>
+                        <td class="rank mono sticky-0" data-label="#">#50</td>
                         <td class="model-name sticky-1" data-label="Модель">mistralai/Ministral-3-3B-Instruct-2512 (Mistral API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1787,7 +1895,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/mistralai/mistral-nemo">
-                        <td class="rank mono sticky-0" data-label="#">#48</td>
+                        <td class="rank mono sticky-0" data-label="#">#51</td>
                         <td class="model-name sticky-1" data-label="Модель">Mistral-Nemo (Mistral API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1823,7 +1931,7 @@
                     </tr>
                     
                     <tr data-model="openai/gpt-4.1-mini">
-                        <td class="rank mono sticky-0" data-label="#">#49</td>
+                        <td class="rank mono sticky-0" data-label="#">#52</td>
                         <td class="model-name sticky-1" data-label="Модель">GPT-4.1-mini</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1859,7 +1967,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/minimax/minimax-m2:free">
-                        <td class="rank mono sticky-0" data-label="#">#50</td>
+                        <td class="rank mono sticky-0" data-label="#">#53</td>
                         <td class="model-name sticky-1" data-label="Модель">MiniMaxAI/MiniMax-M2 (Minimax API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1895,7 +2003,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/minimax/minimax-m2:free">
-                        <td class="rank mono sticky-0" data-label="#">#51</td>
+                        <td class="rank mono sticky-0" data-label="#">#54</td>
                         <td class="model-name sticky-1" data-label="Модель">MiniMaxAI/MiniMax-M2 (Minimax API, recommend params)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1931,7 +2039,7 @@
                     </tr>
                     
                     <tr data-model="openai/gpt-4.1-nano">
-                        <td class="rank mono sticky-0" data-label="#">#52</td>
+                        <td class="rank mono sticky-0" data-label="#">#55</td>
                         <td class="model-name sticky-1" data-label="Модель">GPT-4.1-nano</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -1967,7 +2075,7 @@
                     </tr>
                     
                     <tr data-model="openai/ai-sage/GigaChat3-10B-A1.8B-bf16">
-                        <td class="rank mono sticky-0" data-label="#">#53</td>
+                        <td class="rank mono sticky-0" data-label="#">#56</td>
                         <td class="model-name sticky-1" data-label="Модель">ai-sage/GigaChat3-10B-A1.8B-bf16 (sglang)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -2003,7 +2111,7 @@
                     </tr>
                     
                     <tr data-model="openrouter/openai/gpt-oss-20b">
-                        <td class="rank mono sticky-0" data-label="#">#54</td>
+                        <td class="rank mono sticky-0" data-label="#">#57</td>
                         <td class="model-name sticky-1" data-label="Модель">openai/gpt-oss-20b (Vertex API)</td>
                         <td class="num mono" data-label="Критичные/1000">
                             
@@ -2061,7 +2169,7 @@
         </div>
         <p class="info-text">Если хотите, чтобы я добавил ту или иную модель в лидерборд - не стесняйтесь открыть issue/pull request на Github.</p>
         <p class="info-text">
-            Обновлено: 2026-01-27 11:14:40 | Всего моделей: 54 | <a href="https://github.com/kristaller486/RuQualBench">GitHub</a> | <a href="https://t.me/krists">Telegram</a>
+            Обновлено: 2026-02-02 10:15:23 | Всего моделей: 57 | <a href="https://github.com/kristaller486/RuQualBench">GitHub</a> | <a href="https://t.me/krists">Telegram</a>
         </p>
     </div>
     <script>