--- license: apache-2.0 license_link: https://huggingface.co/Apexintelligence-AI/Apex-35B-V1.0/blob/main/LICENSE library_name: transformers pipeline_tag: text-generation base_model: Qwen/Qwen3.6-35B-A3B base_model_relation: finetune tags: - mixture-of-experts - agent - research ---
| Benchmark | Apex-35B V1.0 · Ours 35B |
Qwen3.6 35B-A3B 35B |
Recursive Flint 39.8B |
GLM 5.3 753B |
GLM 5.3-Flash 320B |
DeepSeek V4.1-Flash-1017 552B + 196B |
|---|---|---|---|---|---|---|
| Research | ||||||
| SciCode Main Problem / Subproblem | 62.6 / 75.6 | 37.5 / 60.6 | 26.6 / 46.0 | 79.2 / 91.3 | 76.6 / 86.8 | 68.8 / 81.9 |
| FrontierScience | 40.0 | 21.7 | 5.0 | 30.7 | 50.0 | 61.7 |
| ASI-Bench | 22.5 | 16.3 | 17.5 | 33.1 | 37.4 | 38.9 |
| BioMysteryBench Human-solvable | 76.7 | 53.4 | 49.3 | 85.6 | 87.7 | 86.3 |
| BioMysteryBench Human-difficult | 35.3 | 11.8 | 5.9 | 39.1 | 41.2 | 29.4 |
| PaperBench | 60.9 | 45.8 | 29.3 | 42.5 | 81.8 | 87.1 |
| Agentic Tasks | ||||||
| BrowseComp | 69.4 | 42.1 | 23.2 | 80.9 | 81.0 | 90.5 |
| WideSearch | 69.9 | 60.9 | 48.3 | 78.4 | 81.1 | 79.8 |
| BFCL V4 | 84.0 | 75.2 | 65.4 | 75.8 | 73.9 | 75.0 |
| AutomationBench | 49.3 | 10.0 | 9.8 | 50.2 | 48.8 | 60.5 |
| APEX-Agents | 64.6 | 41.7 | 33.8 | 68.2 | 73.3 | 74.2 |
| Agents’ Last Exam (ALE) | 19.1 | 12.4 | 19.1 | 28.5 | 26.3 | 35.4 |
| Coding | ||||||
| Terminal-Bench 2.1 | 65.2 | 46.1 | 59.6 | 89.9 | 84.3 | 90.6 |
| SWE-bench Pro v2 | 92.4 | 72.4 | 85.2 | 95.8 | 94.7 | 97.5 |
| SWE-bench Verified No Internet | 73.4 | 63.8 | 74.8 | 79.8 | 72.2 | 82.0 |
| LiveCodeBench v6 | 82.3 | 84.5 | 85.1 | 78.3 | 75.3 | 95.7 |
| SWE-Atlas QnA | 33.1 | 12.1 | 16.1 | 48.7 | 45.2 | 50.8 |
| Reasoning | ||||||
| MMLU-Pro | 87.2 | 79.0 | 70.8 | 88.2 | 85.5 | 68.3 |
| AA-LCR | 77.0 | 74.0 | 59.0 | 80.0 | 79.0 | 87.0 |
| HLE | 47.2 | 36.2 | 18.2 | 61.3 | 55.3 | 42.6 |
| GPQA Diamond | 83.3 | 75.8 | 66.2 | 88.4 | 84.9 | 90.9 |
| Internal Benchmarks | ||||||
| RoboRSI | 44.5 | 35.0 | 30.0 | 58.4 | 52.4 | 74.7 |
| Real2Sim | 21.6 | 2.1 | 2.3 | 25.6 | 21.4 | 36.3 |
| CFD-Bench Wins / Draws / Losses vs. Qwen3.6-35B-A3B | 79 / 0 / 21 | N/A | 54 / 2 / 44 | 99 / 0 / 1 | 98 / 0 / 2 | 100 / 0 / 0 |
| Architecture | Sparse mixture-of-experts (MoE) |
|---|---|
| Parameters | 35B total · 3B active per token |
| Context Length | 262,144 tokens (input + output) |
| Input / Output | Text / Text |
| Thinking Mode | Dynamic |
| Weight Precision | BF16 |
| Model | Download | Parameters | Format |
|---|---|---|---|
| Apex-35B-V1.0 | 🤗 Hugging Face | 35B in total, 3B activated | BF16 safetensors (about 72 GB) |
| Apex-35B-V1.0-GGUF | 🤗 Hugging Face | 35B in total, 3B activated | BF16 (71.1 GB, 2 files) |
| Q8_0 (37.8 GB) | |||
| Q6_K_L (32.8 GB), Q6_K_S (29.9 GB), Q6_K (29.2 GB) | |||
| Q5_K_S (25.6 GB), Q5_K_M (25.3 GB) | |||
| Q4_K_L (24.6 GB), IQ4_NL (22.8 GB), Q4_1 (22.4 GB), Q4_K_M (21.7 GB), Q4_K_S (21.4 GB), Q4_0 (20.4 GB), IQ4_XS (20.4 GB), MXFP4_MOE (20.3 GB) | |||
| IQ3_M (19.4 GB), Q3_K_L (18.7 GB), IQ3_S (18.1 GB), Q3_K_M (17.8 GB), IQ3_XS (16.8 GB), Q3_K_S (16.8 GB), IQ3_XXS (16.1 GB) | |||
| Q2_K (14.3 GB), IQ2_M (13.7 GB), IQ2_S (12.5 GB), IQ2_XS (11.7 GB), IQ2_XXS (11.3 GB) | |||
| IQ1_M (9.9 GB), IQ1_S (9.0 GB) |