enhancement: make deep researcher customizable, flexible and less token expensive

configs/config_cli_default.yml

@@ -18,7 +18,7 @@ general:
 llms:
   nemotron_llm_intent:
     _type: nim
-    model_name: nvidia/nemotron-3-nano-30b-a3b
+    model_name: nvidia/nemotron-3-super-120b-a12b
     base_url: "https://integrate.api.nvidia.com/v1"
     temperature: 0.5
     top_p: 0.9
@@ -27,16 +27,16 @@ llms:
     chat_template_kwargs:
       enable_thinking: true
 
-  nemotron_nano_llm:
+  nemotron_super_llm:
     _type: nim
-    model_name: nvidia/nemotron-3-nano-30b-a3b
+    model_name: nvidia/nemotron-3-super-120b-a12b
     base_url: "https://integrate.api.nvidia.com/v1"
-    temperature: 0.5
-    top_p: 0.9
-    max_tokens: 4096
+    temperature: 0.7
+    top_p: 0.7
+    max_tokens: 65536
     num_retries: 5
     chat_template_kwargs:
-      enable_thinking: false
+      enable_thinking: true
 
   gpt_oss_llm:
     _type: nim
@@ -48,20 +48,6 @@ llms:
     api_key: ${NVIDIA_API_KEY}
     max_retries: 10
 
-  # Nemotron Super is compatible and tested with AIQ but has limited availability
-  # on the Build API due to high demand.
-  # Uncomment nemotron_super_llm below if the endpoint is accessible.
-  # nemotron_super_llm:
-  #   _type: nim
-  #   model_name: nvidia/nemotron-3-super-120b-a12b
-  #   base_url: "https://integrate.api.nvidia.com/v1"
-  #   temperature: 1.0
-  #   top_p: 1.0
-  #   max_tokens: 128000
-  #   num_retries: 5
-  #   chat_template_kwargs:
-  #     enable_thinking: true
-
 functions:
   # =========================================================================
   # Data Source Registry
@@ -122,8 +108,8 @@ functions:
 
   clarifier_agent:
     _type: clarifier_agent
-    llm: nemotron_nano_llm  # replace with nemotron_super_llm if available
-    planner_llm: nemotron_nano_llm  # replace with nemotron_super_llm if available
+    llm: nemotron_super_llm
+    planner_llm: nemotron_super_llm
     # tools: omitted -> inherits all from data_source_registry
     # exclude_tools: []
     max_turns: 3
@@ -133,7 +119,7 @@ functions:
 
   shallow_research_agent:
     _type: shallow_research_agent
-    llm: nemotron_nano_llm
+    llm: nemotron_super_llm
     # tools: omitted -> inherits all from data_source_registry
     exclude_tools:
       - advanced_web_search_tool
@@ -143,9 +129,10 @@ functions:
   deep_research_agent:
     _type: deep_research_agent
     orchestrator_llm: gpt_oss_llm
-    researcher_llm: nemotron_nano_llm  # replace with nemotron_super_llm if available
+    source_router_llm: nemotron_super_llm
+    researcher_llm: nemotron_super_llm
     planner_llm: gpt_oss_llm
-    max_loops: 2
+    writer_llm: gpt_oss_llm
     exclude_tools:
       - web_search_tool
 

configs/config_skills.yml → configs/config_domain_routing_and_skills.yml

@@ -47,24 +47,13 @@ general:
         - "*"
 
 llms:
-  nemotron_llm_intent:
+  nemotron_super_llm:
     _type: nim
-    model_name: nvidia/nemotron-3-nano-30b-a3b
+    model_name: nvidia/nemotron-3-super-120b-a12b
     base_url: "https://integrate.api.nvidia.com/v1"
-    temperature: 0.5
-    top_p: 0.9
-    max_tokens: 4096
-    num_retries: 5
-    chat_template_kwargs:
-      enable_thinking: true
-
-  nemotron_nano_llm:
-    _type: nim
-    model_name: nvidia/nemotron-3-nano-30b-a3b
-    base_url: "https://integrate.api.nvidia.com/v1"
-    temperature: 0.1
-    top_p: 0.3
-    max_tokens: 16384
+    temperature: 0.7
+    top_p: 0.7
+    max_tokens: 65536
     num_retries: 5
     chat_template_kwargs:
       enable_thinking: true
@@ -79,20 +68,6 @@ llms:
     api_key: ${NVIDIA_API_KEY}
     max_retries: 10
 
-  # Nemotron Super is compatible and tested with AIQ but has limited availability
-  # on the Build API due to high demand.
-  # Uncomment nemotron_super_llm below if the endpoint is accessible.
-  # nemotron_super_llm:
-  #   _type: nim
-  #   model_name: nvidia/nemotron-3-super-120b-a12b
-  #   base_url: "https://integrate.api.nvidia.com/v1"
-  #   temperature: 1.0
-  #   top_p: 1.0
-  #   max_tokens: 128000
-  #   num_retries: 5
-  #   chat_template_kwargs:
-  #     enable_thinking: true
-
   # LLM for document summaries (required when generate_summary: true)
   summary_llm:
     _type: nim
@@ -138,17 +113,27 @@ functions:
         tools:
           - web_search_tool
           - advanced_web_search_tool
+      - id: news_search
+        name: "News Search"
+        description: "Search recent news articles and current events."
+        tools:
+          - duckduckgo_news_search_tool
+      - id: prediction_market
+        name: "Prediction Markets"
+        description: "Search Polymarket events and market-implied probabilities."
+        tools:
+          - polymarket_search_tool
       - id: knowledge_layer
         name: "Knowledge Base"
         description: "Search uploaded documents and files."
         tools:
           - knowledge_search
       # Uncomment when paper_search_tool is enabled (requires SERPER_API_KEY)
-      # - id: paper_search
-      #   name: "Academic Papers"
-      #   description: "Search academic papers and scientific publications."
-      #   tools:
-      #     - paper_search_tool
+      - id: paper_search
+        name: "Academic Papers"
+        description: "Search academic papers and scientific publications."
+        tools:
+          - paper_search_tool
 
   web_search_tool:
     _type: tavily_web_search
@@ -160,6 +145,15 @@ functions:
     max_results: 2
     advanced_search: true
 
+  duckduckgo_news_search_tool:
+    _type: duckduckgo_news_search
+    max_results: 5
+    timelimit: w
+
+  polymarket_search_tool:
+    _type: polymarket_search
+    max_results: 5
+
   # Knowledge Retrieval (see sources/knowledge_layer/KNOWLEDGE-LAYER-SETUP.md)
   knowledge_search:
     _type: knowledge_retrieval
@@ -173,10 +167,10 @@ functions:
 
   # Paper Search (optional - requires SERPER_API_KEY)
   # Uncomment the block below and set SERPER_API_KEY to enable academic paper search.
-  # paper_search_tool:
-  #   _type: paper_search
-  #   max_results: 5
-  #   serper_api_key: ${SERPER_API_KEY}
+  paper_search_tool:
+    _type: paper_search
+    max_results: 5
+    serper_api_key: ${SERPER_API_KEY}
 
   # =========================================================================
   # Agents
@@ -186,46 +180,28 @@ functions:
   # from an agent (e.g., give shallow the basic search, deep the advanced).
   # To bypass auto-inherit entirely, set an explicit `tools` list.
   # =========================================================================
-  intent_classifier:
-    _type: intent_classifier
-    llm: nemotron_llm_intent
-    # tools: omitted -> inherits all from data_source_registry
-    # exclude_tools: []
-    verbose: true
-
-  clarifier_agent:
-    _type: clarifier_agent
-    llm: nemotron_nano_llm  # replace with nemotron_super_llm if available
-    planner_llm: nemotron_nano_llm  # replace with nemotron_super_llm if available
-    # tools: omitted -> inherits all from data_source_registry
-    # exclude_tools: []
-    max_turns: 3
-    enable_plan_approval: true
-    log_response_max_chars: 2000
-    verbose: true
-
-  shallow_research_agent:
-    _type: shallow_research_agent
-    llm: nemotron_nano_llm
-    # tools: omitted -> inherits all from data_source_registry
-    exclude_tools:                     # Remove advanced variant; shallow uses web_search_tool
-      - advanced_web_search_tool
-    verbose: true
-    max_llm_turns: 10
-    max_tool_iterations: 5
 
   deep_research_agent:
     _type: deep_research_agent
     orchestrator_llm: gpt_oss_llm
-    researcher_llm: nemotron_nano_llm  # replace with nemotron_super_llm if available
+    source_router_llm: nemotron_super_llm
+    researcher_llm: nemotron_super_llm
     planner_llm: gpt_oss_llm
+    writer_llm: gpt_oss_llm
     # tools: omitted -> inherits all from data_source_registry
     exclude_tools:                     # Remove basic variant; deep uses advanced_web_search_tool
       - web_search_tool
-    max_loops: 2
     verbose: true
+    domain_catalog_path: configs/domain_catalogs/deep_research_domain_catalog.yml
     skills:
       enabled: true
+      agent_sources:
+        researcher:
+          - /skills/research-sandbox/
+        writer-agent:
+          - /skills/synthesis/
+      sandbox_required_sources:
+        - /skills/research-sandbox/
     sandbox:
       provider: modal
       app_name: aiq-deep-research
@@ -239,9 +215,6 @@ functions:
       block_network: true
 
 workflow:
-  _type: chat_deepresearcher_agent
+  _type: deep_research_workflow
   verbose: true
-  enable_escalation: true
-  enable_clarifier: true
   use_async_deep_research: true
-  checkpoint_db: ${AIQ_CHECKPOINT_DB:-./checkpoints.db}

configs/config_frontier_models.yml

@@ -50,7 +50,7 @@ general:
 llms:
   nemotron_llm_intent:
     _type: nim
-    model_name: nvidia/nemotron-3-nano-30b-a3b
+    model_name: nvidia/nemotron-3-super-120b-a12b
     base_url: "https://integrate.api.nvidia.com/v1"
     temperature: 0.5
     top_p: 0.9
@@ -59,9 +59,9 @@ llms:
     chat_template_kwargs:
       enable_thinking: true
 
-  nemotron_nano_llm:
+  nemotron_super_llm:
     _type: nim
-    model_name: nvidia/nemotron-3-nano-30b-a3b
+    model_name: nvidia/nemotron-3-super-120b-a12b
     base_url: "https://integrate.api.nvidia.com/v1"
     temperature: 0.1
     top_p: 0.3
@@ -77,20 +77,6 @@ llms:
     model_name: "gpt-5.2"
     api_key: ${OPENAI_API_KEY}
 
-  # Nemotron Super is compatible and tested with AIQ but has limited availability
-  # on the Build API due to high demand.
-  # Uncomment nemotron_super_llm below if the endpoint is accessible.
-  # nemotron_super_llm:
-  #   _type: nim
-  #   model_name: nvidia/nemotron-3-super-120b-a12b
-  #   base_url: "https://integrate.api.nvidia.com/v1"
-  #   temperature: 1.0
-  #   top_p: 1.0
-  #   max_tokens: 128000
-  #   num_retries: 5
-  #   chat_template_kwargs:
-  #     enable_thinking: true
-
   # LLM for document summaries (required when generate_summary: true)
   summary_llm:
     _type: nim
@@ -152,7 +138,7 @@ functions:
 
   shallow_research_agent:
     _type: shallow_research_agent
-    llm: nemotron_nano_llm
+    llm: nemotron_super_llm
     verbose: true
     tools:
       - web_search_tool
@@ -163,9 +149,10 @@ functions:
   deep_research_agent:
     _type: deep_research_agent
     orchestrator_llm: openai_gpt_5_2
-    researcher_llm: nemotron_nano_llm  # replace with nemotron_super_llm if available
+    source_router_llm: nemotron_super_llm
+    researcher_llm: nemotron_super_llm
     planner_llm: openai_gpt_5_2
-    max_loops: 2
+    writer_llm: openai_gpt_5_2
     verbose: true
     tools:
       # - paper_search_tool  # Uncomment if SERPER_API_KEY is set