x_ai.yml

# yaml-language-server: $schema=.schema.json
name: X AI
id: x-ai
pricing_urls:
  - https://docs.x.ai/docs/models
api_pattern: 'https://api\.x\.ai'
model_match:
  contains: grok
provider_match:
  equals: xai
extractors:
  # xAI native SDK format (xai-sdk) — protobuf SamplingUsage fields
  # https://github.com/xai-org/xai-sdk-python/blob/main/src/xai_sdk/proto/v6/usage_pb2.pyi
  # xAI reports completion and reasoning as separate output counts, while both are billed at the output rate.
  # https://docs.x.ai/developers/tools/tool-usage-details#understanding-token-usage
  - api_flavor: default
    root: usage
    mappings:
      - path: prompt_tokens
        dest: input_tokens
      - path: cached_prompt_text_tokens
        dest: cache_read_tokens
        required: false
      - path: reasoning_tokens
        dest: output_tokens
        required: false
      - path: reasoning_tokens
        dest: output_reasoning_tokens
        required: false
      - path: completion_tokens
        dest: output_tokens
  # OpenAI-compatible chat completions format
  - api_flavor: chat
    root: usage
    mappings:
      - path: prompt_tokens
        dest: input_tokens
      - path: [prompt_tokens_details, cached_tokens]
        dest: cache_read_tokens
        required: false
      - path: [completion_tokens_details, audio_tokens]
        dest: output_audio_tokens
        required: false
      - path: [completion_tokens_details, reasoning_tokens]
        dest: output_tokens
        required: false
      - path: [completion_tokens_details, reasoning_tokens]
        dest: output_reasoning_tokens
        required: false
      - path: completion_tokens
        dest: output_tokens

models:
  - id: grok-2-1212
    name: Grok 2 1212
    description: >-
      (deprecated) Grok 2 1212 introduces significant enhancements to accuracy, instruction adherence, and multilingual support,
      making
      it a powerful and flexible choice for developers seeking a highly steerable, intelligent model.
    match:
      or:
        - equals: grok-2-1212
        - equals: grok-2
        - equals: grok-2-latest
    context_window: 32768
    deprecated: true
    prices_checked: 2025-07-12
    prices:
      input_mtok: 2
      output_mtok: 10

  - id: grok-2-vision-1212
    name: Grok 2 Vision 1212
    description: Our multimodal model that processes documents, diagrams, charts, screenshots, and photographs.
    match:
      or:
        - equals: grok-2-vision-1212
        - equals: grok-2-vision
        - equals: grok-2-vision-latest
    context_window: 32768
    prices_checked: 2025-07-12
    prices:
      input_mtok: 2
      output_mtok: 10

  - id: grok-3
    name: Grok 3
    description: >-
      Flagship model that excels at enterprise use cases like data extraction,
      coding, and text summarization. Possesses deep domain knowledge in finance, healthcare, law, and science.
    match:
      or:
        - equals: grok-3
        - equals: grok-3-latest
        - equals: grok-3-beta
    context_window: 131072
    prices_checked: 2025-07-12
    prices:
      input_mtok: 3
      cache_read_mtok: 0.75
      output_mtok: 15

  - id: grok-3-fast
    name: Grok 3 Fast
    description: >-
      Excels at enterprise use cases like data extraction, coding, and text summarization. Possesses deep domain knowledge
      in finance, healthcare, law, and science.
    match:
      or:
        - equals: grok-3-fast
        - equals: grok-3-fast-latest
        - equals: grok-3-fast-beta
    context_window: 131072
    prices_checked: 2025-07-12
    prices:
      input_mtok: 5
      cache_read_mtok: 1.25
      output_mtok: 25

  - id: grok-3-mini
    name: Grok 3 Mini
    description: >-
      A lightweight model that thinks before responding. Fast, smart, and great for logic-based tasks that do not require
      deep domain knowledge. The raw thinking traces are accessible.
    match:
      or:
        - equals: grok-3-mini
        - equals: grok-3-mini-beta
        - equals: grok-3-mini-latest
    context_window: 131072
    prices_checked: 2025-07-12
    prices:
      input_mtok: 0.3
      cache_read_mtok: 0.075
      output_mtok: 0.5

  - id: grok-3-mini-fast
    name: Grok 3 Mini Fast
    description: >-
      A lightweight model that thinks before responding. Fast, smart, and great for logic-based tasks that do not require
      deep domain knowledge. The raw thinking traces are accessible.
    match:
      or:
        - equals: grok-3-mini-fast
        - equals: grok-3-mini-fast-beta
        - equals: grok-3-mini-fast-latest
    context_window: 131072
    prices_checked: 2025-07-12
    prices:
      input_mtok: 0.6
      cache_read_mtok: 0.15
      output_mtok: 4

  - id: grok-4-0709
    name: Grok 4
    description: >-
      A flagship model, offering unparalleled performance in natural language, math and reasoning - the perfect jack of all
      trades.
    match:
      or:
        - equals: grok-4-0709
        - equals: grok-4
        - equals: grok-4-latest
    context_window: 256000
    prices_checked: 2025-07-12
    prices:
      input_mtok: 3
      cache_read_mtok: 0.75
      output_mtok: 15

  - id: grok-4-1-fast-non-reasoning
    name: Grok 4.1 Fast Non-Reasoning
    description: >-
      A frontier multimodal model optimized specifically for high-performance agentic tool calling.
    match:
      or:
        - equals: grok-4-1-fast-non-reasoning
        - equals: grok-4-1-fast-non-reasoning-latest
    context_window: 2000000
    prices_checked: 2026-01-19
    prices:
      input_mtok: 0.2
      cache_read_mtok: 0.05
      output_mtok: 0.5

  - id: grok-4-1-fast-reasoning
    name: Grok 4.1 Fast Reasoning
    description: >-
      A frontier multimodal model optimized specifically for high-performance agentic tool calling.
    match:
      or:
        - equals: grok-4-1-fast
        - equals: grok-4-1-fast-reasoning
        - equals: grok-4-1-fast-reasoning-latest
    context_window: 2000000
    prices_checked: 2026-01-19
    prices:
      input_mtok: 0.2
      cache_read_mtok: 0.05
      output_mtok: 0.5

  - id: grok-4-fast-non-reasoning
    name: Grok 4 Fast Non-Reasoning
    description: >-
      A frontier multimodal model optimized specifically for high-performance agentic tool calling.
    match:
      or:
        - equals: grok-4-fast-non-reasoning
        - equals: grok-4-fast-non-reasoning-latest
    context_window: 2000000
    prices_checked: 2026-01-19
    prices:
      input_mtok: 0.2
      cache_read_mtok: 0.05
      output_mtok: 0.5

  - id: grok-4-fast-reasoning
    name: Grok 4 Fast Reasoning
    description: >-
      A frontier multimodal model optimized specifically for high-performance agentic tool calling.
    match:
      or:
        - equals: grok-4-fast
        - equals: grok-4-fast-reasoning
        - equals: grok-4-fast-reasoning-latest
    context_window: 2000000
    prices_checked: 2026-01-19
    prices:
      input_mtok: 0.2
      cache_read_mtok: 0.05
      output_mtok: 0.5

  - id: grok-4.20
    name: Grok 4.20
    description: >-
      Grok 4.20 is a reasoning model from xAI with industry-leading speed and agentic tool calling capabilities. It combines
      low hallucination rates with strict prompt adherence.
    match:
      equals: grok-4.20
    prices_checked: 2026-06-09
    prices:
      input_mtok: 1.25
      cache_read_mtok: 0.2
      output_mtok: 2.5

  - id: grok-4.20-multi-agent
    name: Grok 4.20 Multi-Agent
    description: >-
      Grok 4.20 Multi-Agent is a variant of xAI's Grok 4.20 designed for collaborative, agent-based workflows. Multiple agents
      operate in parallel to conduct deep research, coordinate tool use, and synthesize information.
    match:
      equals: grok-4.20-multi-agent
    prices_checked: 2026-06-09
    prices:
      input_mtok: 2
      cache_read_mtok: 0.2
      output_mtok: 6

  - id: grok-4.3
    name: Grok 4.3
    description: >-
      Most advanced flagship model, leading the industry in non-hallucination rate, agentic tool calling,
      and instruction following capabilities. Supports text and image inputs with text outputs, function
      calling, structured outputs, and reasoning.
    match:
      or:
        - equals: grok-4.3
        - regex: ^grok-4\.3-\d{8}$
        - equals: x-ai/grok-4.3
        - regex: ^x-ai/grok-4\.3-\d{8}$
        - equals: grok-4.3-latest
    context_window: 1000000
    prices_checked: 2026-05-12
    prices:
      input_mtok: 1.25
      cache_read_mtok: 0.2
      output_mtok: 2.5

  - id: grok-4.5
    name: Grok 4.5
    description: >-
      xAI's most intelligent and fastest flagship model, well-suited for general-purpose use including coding and chat.
      Supports text and image inputs with text outputs, function calling, structured outputs, and reasoning.
    match:
      or:
        - equals: grok-4.5
        - regex: ^grok-4\.5-\d{8}$
        - equals: x-ai/grok-4.5
        - regex: ^x-ai/grok-4\.5-\d{8}$
        - equals: grok-4.5-latest
        - equals: grok-latest
    context_window: 500000
    prices_checked: 2026-07-08
    prices:
      input_mtok: 2
      cache_read_mtok: 0.5
      output_mtok: 6

  - id: grok-build-0.1
    name: Grok Build 0.1
    description: >-
      Grok Build 0.1 is xAI's fast coding model trained specifically for agentic software engineering workflows. It supports
      text and image inputs with text output, and is optimized for interactive coding.
    match:
      equals: grok-build-0.1
    prices_checked: 2026-06-09
    prices:
      input_mtok: 1
      cache_read_mtok: 0.2
      output_mtok: 2

  - id: grok-code-fast-1
    name: Grok Code Fast 1
    description: >-
      A speedy and economical reasoning model that excels at agentic coding.
    match:
      or:
        - equals: grok-code-fast
        - equals: grok-code-fast-1
        - equals: grok-code-fast-1-0825
    context_window: 256000
    prices_checked: 2026-01-19
    prices:
      input_mtok: 0.2
      cache_read_mtok: 0.02
      output_mtok: 1.5