inclusionAI
/

LLaDA2.0-mini-preview

@@ -20,11 +20,10 @@ tags:
 | Benchmark | Ling-mini-2.0 | LLaDA-MoE-7B-A1B-Instruct | LLaDA2.0-mini-preview |
 | :------------------------------ | :-------------: | :-------------------------: | :---------------------: |
-| **Average** | 64.31 | 53.73 | 60.98 |
 | **Knowledge** | | | |
 | MMLU | 78.75 | 67.18 | 72.49 |
 | MMLU-PRO | 56.40 | 44.64 | 49.22 |
-| GPQA | 37.99 | 31.09 | 31.82 |
 | CMMLU | 77.84 | 64.30 | 67.53 |
 | C-EVAL | 77.85 | 63.93 | 66.54 |
 | **Reasoning** | | | |
@@ -36,12 +35,10 @@ tags:
 | mbpp | 81.03 | 70.02 | 77.75 |
 | MultiPL-E | 62.23 | 52.53 | 62.43 |
 | humaneval | 77.44 | 61.59 | 80.49 |
-| livecodebench_v6 | 30.18 | 13.27 | 19.93 |
 | Bigcodebench-Full | 35.88 | 20.44 | 30.44 |
 | **Math** | | | |
 | GSM8K | 91.58 | 82.41 | 89.01 |
 | math | 82.22 | 58.68 | 73.50 |
-| OlympiadBench | 49.93 | 21.04 | 36.67 |
 | **Agent & Alignment** | | | |
 | BFCL_Live | 45.74 | 63.09 | 74.11 |
 | IFEval-strict -prompt | 69.13 | 59.33 | 62.50 |
@@ -60,12 +57,18 @@ Supports **tool calling** and achieves excellent performance in complex agent-ba
 + **Open & Extensible**:
 Fully open-source with commitment to transparency. We plan to release a **leading inference framework** in the future and continue investing in cutting-edge areas like **diffusion LLMs (dLLM)** to drive disruptive innovation.
 ---
 ## 📦 Model Variants
 | Model ID | Description | Hugging Face Link |
 | --- | --- | --- |
 | `inclusionAI/LLaDA2.0-mini-preview` | Instruction-tuned model, ready for downstream applications. | [🤗 Model Card](https://huggingface.co/inclusionAI/LLaDA2.0-mini-preview) |
 ---

 | Benchmark | Ling-mini-2.0 | LLaDA-MoE-7B-A1B-Instruct | LLaDA2.0-mini-preview |
 | :------------------------------ | :-------------: | :-------------------------: | :---------------------: |
+| **Average** | 68.98 | 59.72 | 66.89 |
 | **Knowledge** | | | |
 | MMLU | 78.75 | 67.18 | 72.49 |
 | MMLU-PRO | 56.40 | 44.64 | 49.22 |
 | CMMLU | 77.84 | 64.30 | 67.53 |
 | C-EVAL | 77.85 | 63.93 | 66.54 |
 | **Reasoning** | | | |
 | mbpp | 81.03 | 70.02 | 77.75 |
 | MultiPL-E | 62.23 | 52.53 | 62.43 |
 | humaneval | 77.44 | 61.59 | 80.49 |
 | Bigcodebench-Full | 35.88 | 20.44 | 30.44 |
 | **Math** | | | |
 | GSM8K | 91.58 | 82.41 | 89.01 |
 | math | 82.22 | 58.68 | 73.50 |
 | **Agent & Alignment** | | | |
 | BFCL_Live | 45.74 | 63.09 | 74.11 |
 | IFEval-strict -prompt | 69.13 | 59.33 | 62.50 |
 + **Open & Extensible**:
 Fully open-source with commitment to transparency. We plan to release a **leading inference framework** in the future and continue investing in cutting-edge areas like **diffusion LLMs (dLLM)** to drive disruptive innovation.
+## 🗺️ What's Next
++ **Supercharged Reasoning with LLaDA 2.0:** LLaDA 2.0 series will be fine-tuned with **Reinforcement Learning**, unlocking a new level of sophisticated reasoning and problem-solving abilities.
++ **Tools for Innovators:** we will release a **detailed tutorial** and our complete **post-training framework**. Whether you want to master the current model or build your own customized versions, you'll have the tools you need. Stay tuned
 ---
 ## 📦 Model Variants
 | Model ID | Description | Hugging Face Link |
 | --- | --- | --- |
 | `inclusionAI/LLaDA2.0-mini-preview` | Instruction-tuned model, ready for downstream applications. | [🤗 Model Card](https://huggingface.co/inclusionAI/LLaDA2.0-mini-preview) |
+| `inclusionAI/LLaDA2.0-flash-preview` | Instruction-tuned model, ready for downstream applications. | [🤗 Model Card](https://huggingface.co/inclusionAI/LLaDA2.0-flash-preview) |
 ---