Tongyi DeepResearch 30B A3B

Description

Tongyi DeepResearch is an agentic large language model developed by Tongyi Lab, with 30 billion total parameters activating only 3 billion per token. It's optimized for long-horizon, deep information-seeking tasks and delivers state-of-the-art performance on benchmarks like Humanity's Last Exam, BrowserComp, BrowserComp-ZH, WebWalkerQA, GAIA, xbench-DeepSearch, and FRAMES. This makes it superior for complex agentic search, reasoning, and multi-step problem-solving compared to prior models. The model includes a fully automated synthetic data pipeline for scalable pre-training, fine-tuning, and reinforcement learning. It uses large-scale continual pre-training on diverse agentic data to boost reasoning and stay fresh. It also features end-to-end on-policy RL with a customized Group Relative Policy Optimization, including token-level gradients and negative sample filtering for stable training. The model supports ReAct for core ability checks and an IterResearch-based 'Heavy' mode for max performance through test-time scaling. It's ideal for advanced research agents, tool use, and heavy inference workflows.

Raw fields snapshot

{
  "id": "alibaba/tongyi-deepresearch-30b-a3b",
  "name": "Tongyi DeepResearch 30B A3B",
  "description": "Tongyi DeepResearch is an agentic large language model developed by Tongyi Lab, with 30 billion total parameters activating only 3 billion per token. It's optimized for long-horizon, deep information-seeking tasks and delivers state-of-the-art performance on benchmarks like Humanity's Last Exam, BrowserComp, BrowserComp-ZH, WebWalkerQA, GAIA, xbench-DeepSearch, and FRAMES. This makes it superior for complex agentic search, reasoning, and multi-step problem-solving compared to prior models.\n\nThe model includes a fully automated synthetic data pipeline for scalable pre-training, fine-tuning, and reinforcement learning. It uses large-scale continual pre-training on diverse agentic data to boost reasoning and stay fresh. It also features end-to-end on-policy RL with a customized Group Relative Policy Optimization, including token-level gradients and negative sample filtering for stable training. The model supports ReAct for core ability checks and an IterResearch-based 'Heavy' mode for max performance through test-time scaling. It's ideal for advanced research agents, tool use, and heavy inference workflows.",
  "created": 1758210804,
  "canonical_slug": "alibaba/tongyi-deepresearch-30b-a3b",
  "hugging_face_id": "Alibaba-NLP/Tongyi-DeepResearch-30B-A3B",
  "source_type": "openrouter_only",
  "context_length": 131072,
  "max_completion_tokens": 131072,
  "is_moderated": false,
  "architecture": {
    "modality": "text->text",
    "input_modalities": [
      "text"
    ],
    "output_modalities": [
      "text"
    ],
    "tokenizer": "Other",
    "instruct_type": null
  },
  "input_modalities": [
    "text"
  ],
  "output_modalities": [
    "text"
  ],
  "modality": "text->text",
  "tokenizer": "Other",
  "instruct_type": null,
  "supported_parameters": [
    "include_reasoning",
    "max_tokens",
    "reasoning",
    "response_format",
    "structured_outputs",
    "temperature",
    "tool_choice",
    "tools",
    "top_p"
  ],
  "default_parameters": {
    "temperature": null,
    "top_p": null,
    "frequency_penalty": null
  },
  "per_request_limits": null,
  "top_provider": {
    "context_length": 131072,
    "max_completion_tokens": 131072,
    "is_moderated": false
  },
  "pricing": {
    "prompt": "0.00000009",
    "completion": "0.00000045",
    "input_cache_read": "0.00000009"
  },
  "PPM": {
    "prompt": 0.09,
    "completion": 0.45,
    "input_cache_read": 0.09
  },
  "openrouter_raw": {
    "id": "alibaba/tongyi-deepresearch-30b-a3b",
    "canonical_slug": "alibaba/tongyi-deepresearch-30b-a3b",
    "hugging_face_id": "Alibaba-NLP/Tongyi-DeepResearch-30B-A3B",
    "name": "Tongyi DeepResearch 30B A3B",
    "created": 1758210804,
    "description": "Tongyi DeepResearch is an agentic large language model developed by Tongyi Lab, with 30 billion total parameters activating only 3 billion per token. It's optimized for long-horizon, deep information-seeking tasks and delivers state-of-the-art performance on benchmarks like Humanity's Last Exam, BrowserComp, BrowserComp-ZH, WebWalkerQA, GAIA, xbench-DeepSearch, and FRAMES. This makes it superior for complex agentic search, reasoning, and multi-step problem-solving compared to prior models.\n\nThe model includes a fully automated synthetic data pipeline for scalable pre-training, fine-tuning, and reinforcement learning. It uses large-scale continual pre-training on diverse agentic data to boost reasoning and stay fresh. It also features end-to-end on-policy RL with a customized Group Relative Policy Optimization, including token-level gradients and negative sample filtering for stable training. The model supports ReAct for core ability checks and an IterResearch-based 'Heavy' mode for max performance through test-time scaling. It's ideal for advanced research agents, tool use, and heavy inference workflows.",
    "context_length": 131072,
    "architecture": {
      "modality": "text->text",
      "input_modalities": [
        "text"
      ],
      "output_modalities": [
        "text"
      ],
      "tokenizer": "Other",
      "instruct_type": null
    },
    "pricing": {
      "prompt": "0.00000009",
      "completion": "0.00000045",
      "input_cache_read": "0.00000009"
    },
    "top_provider": {
      "context_length": 131072,
      "max_completion_tokens": 131072,
      "is_moderated": false
    },
    "per_request_limits": null,
    "supported_parameters": [
      "include_reasoning",
      "max_tokens",
      "reasoning",
      "response_format",
      "structured_outputs",
      "temperature",
      "tool_choice",
      "tools",
      "top_p"
    ],
    "default_parameters": {
      "temperature": null,
      "top_p": null,
      "frequency_penalty": null
    },
    "expiration_date": null
  }
}