---
title: Grok 4.3 | Runware Docs
url: https://runware.ai/docs/models/xai-grok-4-3
description: Flagship multimodal LLM for agentic tool calling, long context, and low-hallucination reasoning
---
# Grok 4.3

![Grok 4.3](https://assets.runware.ai/covers/xai-grok-4-3.jpg)

Grok 4.3 is xAI's flagship language model for agentic reasoning, strong instruction following, and minimal hallucinations. It supports text and image input, a 1 million token context window, configurable reasoning effort including non-reasoning mode, function calling, and structured outputs for production assistants, coding workflows, and long-context analysis.

- **Status**: live
- **ID**: `xai-grok-4-3`
- Try in Playground
- **Context window**: 1M context
- **Max output**: 128K output
- **Creator**: xAI
- **Release Date**: May 5, 2026
- **Capabilities**: Text to Text, Image to Text, Checkpoint

## Pricing

~$0.00144 typical run

**Rates** (per 1M tokens)

- Input: `$1.25`
- Output: `$2.50`
- Cached input: `$0.20`

**4 example generations**

- [tool-calling](https://runware.ai/docs/models/xai-grok-4-3/examples#example-newsroom-embargo-timestamp-lookup): `$0.000206` (~4s)
- [tool-calling](https://runware.ai/docs/models/xai-grok-4-3/examples#example-pharmaceutical-recall-timestamp-record): `$0.000236` (~4s)
- [text-generation](https://runware.ai/docs/models/xai-grok-4-3/examples#example-saas-pricing-migration-decision-memo): `$0.00265` (~15s)
- [text-generation](https://runware.ai/docs/models/xai-grok-4-3/examples#example-clinical-trial-consent-rewrite): `$0.00313` (~12s)

billed per token, so a request costs what the model produced · times vary, measured end to end · [How pricing works](https://runware.ai/docs/models-api/pricing)

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/xai-grok-4-3#request-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `textInference`

Identifier for the type of task being performed.

### [taskUUID](https://runware.ai/docs/models/xai-grok-4-3#request-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier for tracking tasks and matching async responses. Must be unique per task.

### [outputFormat](https://runware.ai/docs/models/xai-grok-4-3#request-outputformat)

- **Type**: `string`
- **Default**: `TEXT`

Specifies the file format of the generated output. The available values depend on the task type and the specific model's capabilities.

**Allowed values**: `TEXT`

### [webhookURL](https://runware.ai/docs/models/xai-grok-4-3#request-webhookurl)

- **Type**: `string`
- **Format**: `uri`

Webhook URL that receives JSON responses via HTTP POST when generation tasks complete. For batch requests with multiple results, each completed item triggers a separate webhook call as it becomes available.

**Learn more** (1 resource):

- [Webhooks](https://runware.ai/docs/models-api/webhooks) (platform)

### [deliveryMethod](https://runware.ai/docs/models/xai-grok-4-3#request-deliverymethod)

- **Type**: `string`
- **Default**: `sync`

Determines how the API delivers task results.

**Allowed values**:

- `sync` Returns results directly in the API response. Over HTTP, a task still running after 120 seconds continues as an async task and comes back with status 'processing'. Poll for its results using getResponse.
- `async` Returns an immediate acknowledgment with the task UUID. Poll for results using getResponse.
- `stream` Streams results token-by-token as they are generated.

**Learn more** (1 resource):

- [Task Polling](https://runware.ai/docs/models-api/task-polling) (platform)

### [includeCost](https://runware.ai/docs/models/xai-grok-4-3#request-includecost)

- **Type**: `boolean`

Include task cost in the response.

### [includeUsage](https://runware.ai/docs/models/xai-grok-4-3#request-includeusage)

- **Type**: `boolean`

Include token usage statistics in the response.

### [numberResults](https://runware.ai/docs/models/xai-grok-4-3#request-numberresults)

- **Type**: `integer`
- **Min**: `1`
- **Max**: `4`
- **Default**: `1`

Number of results to generate. Each result uses a different seed, producing variations of the same parameters.

**Inputs**

Input resources for the task (images, audio, etc). These must be nested inside the \`inputs\` object.

### [images](https://runware.ai/docs/models/xai-grok-4-3#request-inputs-images)

- **Path**: `inputs.images`
- **Type**: `array of strings`
- **Min items**: `1`

Array of image inputs (UUID, URL, Data URI, or Base64).

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/xai-grok-4-3#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `xai-grok-4-3`

Send `auto` instead of an AIR identifier to have the model router choose the model for this task. The rest of the request is unchanged, every value you send is kept, and only the parameters you leave out are filled in. See [Model Router](https://runware.ai/docs/models-api/model-router).

**Learn more** (3 resources):

- [Text To Image: Model Selection](https://runware.ai/docs/learn/text-to-image#model-selection) (learn)
- [Image Inpainting: Model Specialized Inpainting Models](https://runware.ai/docs/learn/image-inpainting#model-specialized-inpainting-models) (learn)
- [Image Outpainting: Other Critical Parameters](https://runware.ai/docs/learn/image-outpainting#other-critical-parameters) (learn)

### [seed](https://runware.ai/docs/models/xai-grok-4-3#request-seed)

- **Type**: `integer`
- **Min**: `0`
- **Max**: `9223372036854776000`

Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range.

**Learn more** (1 resource):

- [Seed](https://runware.ai/docs/learn/seed) (learn)

### [messages](https://runware.ai/docs/models/xai-grok-4-3#request-messages)

- **Path**: `messages.role`
- **Type**: `array of objects (2 properties)`
- **Required**: true
- **Min items**: `1`

Array of chat messages forming the conversation context. The final message must use the user role.

#### [role](https://runware.ai/docs/models/xai-grok-4-3#request-messages-role)

- **Path**: `messages.role`
- **Type**: `string`
- **Required**: true

Role of the message author.

**Allowed values**: `user` `assistant`

#### [content](https://runware.ai/docs/models/xai-grok-4-3#request-messages-content)

- **Path**: `messages.content`
- **Type**: `string`
- **Required**: true
- **Min**: `1`

Text content of the message.

**Settings**

Technical parameters to fine-tune the inference process. These must be nested inside the \`settings\` object.

### [temperature](https://runware.ai/docs/models/xai-grok-4-3#request-settings-temperature)

- **Path**: `settings.temperature`
- **Type**: `float`
- **Min**: `0`
- **Max**: `2`
- **Step**: `0.01`
- **Default**: `1`

Controls randomness in generation. Lower values produce more deterministic outputs, higher values increase variation and creativity.

### [topP](https://runware.ai/docs/models/xai-grok-4-3#request-settings-topp)

- **Path**: `settings.topP`
- **Type**: `float`
- **Min**: `0`
- **Max**: `1`
- **Step**: `0.01`
- **Default**: `1`

Nucleus sampling parameter that controls diversity by limiting the probability mass. Lower values make outputs more focused, higher values increase diversity.

### [maxTokens](https://runware.ai/docs/models/xai-grok-4-3#request-settings-maxtokens)

- **Path**: `settings.maxTokens`
- **Type**: `integer`
- **Min**: `1`

Maximum number of tokens to generate in the response.

### [thinkingLevel](https://runware.ai/docs/models/xai-grok-4-3#request-settings-thinkinglevel)

- **Path**: `settings.thinkingLevel`
- **Type**: `string`
- **Default**: `low`

Controls the depth of internal reasoning the model performs before generating a response.

**Allowed values**: `none` `low` `medium` `high`

### [toolChoice](https://runware.ai/docs/models/xai-grok-4-3#request-toolchoice)

- **Path**: `toolChoice.type`
- **Type**: `object (2 properties)`

Controls how the model selects which tool to call. This only takes effect when `tools` are defined.

**Examples**:

**Let the model decide (default):**

```json
"toolChoice": {
  "type": "auto"
}
```

**Force a specific tool call:**

```json
"toolChoice": {
  "type": "tool",
  "name": "get_weather"
}
```

**Require any tool call:**

```json
"toolChoice": {
  "type": "any"
}
```

#### [type](https://runware.ai/docs/models/xai-grok-4-3#request-toolchoice-type)

- **Path**: `toolChoice.type`
- **Type**: `string`
- **Required**: true

Strategy the model uses to decide when and which tools to call.

**Allowed values**:

- `auto` Tool calls decided by the model from the conversation context. This is the recommended default.
- `any` At least one tool call, with the model choosing which tool. Useful when you always need structured output.
- `tool` Call to the specific tool identified by `name`. Use this to force a particular function call.
- `none` No tool calls, even if tools are defined. Useful for forcing a text-only response.

#### [name](https://runware.ai/docs/models/xai-grok-4-3#request-toolchoice-name)

- **Path**: `toolChoice.name`
- **Type**: `string`

Name of the specific tool the model must call. Required when type is `tool`.

### [tools](https://runware.ai/docs/models/xai-grok-4-3#request-tools)

- **Path**: `tools.type`
- **Type**: `array of objects (4 properties)`
- **Min items**: `1`
- **Max items**: `128`

An array of tool definitions that the model may call during generation. The model can invoke one or more tools based on the conversation context, outputting structured calls with arguments instead of (or alongside) free-text.

For `function` tools, each definition requires:

- **`type`**: `"function"`
- **`name`**: Unique identifier (alphanumeric, hyphens, underscores; max 64 chars).
- **`description`**: What the function does. The model uses this to decide when to call it.
- **`schema`**: JSON Schema object describing the expected input arguments.

> [!NOTE]
> The `search` tool is executed server-side by the provider. You don't need to handle the tool result yourself.

> [!NOTE]
> The `codeInterpreter` tool is executed server-side by the provider. You don't need to handle the tool result yourself.

**Examples**:

**Function tool, weather lookup:**

```json
"tools": [
  {
    "type": "function",
    "name": "get_weather",
    "description": "Get current weather for a city",
    "schema": {
      "type": "object",
      "properties": {
        "city": { "type": "string", "description": "City name" }
      },
      "required": ["city"]
    }
  }
],
"toolChoice": { "type": "auto" }
```

**Built-in web search:**

```json
"tools": [
  { "type": "search" }
]
```

**Built-in code interpreter:**

```json
"tools": [
  { "type": "codeInterpreter" }
]
```

**Multiple function tools:**

```json
"tools": [
  {
    "type": "function",
    "name": "search_products",
    "description": "Search the product catalog by query and filters.",
    "schema": {
      "type": "object",
      "properties": {
        "query": { "type": "string" },
        "category": { "type": "string" }
      },
      "required": ["query"]
    }
  },
  {
    "type": "function",
    "name": "add_to_cart",
    "description": "Add a product to the user's shopping cart.",
    "schema": {
      "type": "object",
      "properties": {
        "productId": { "type": "string" },
        "quantity": { "type": "integer", "minimum": 1 }
      },
      "required": ["productId"]
    }
  }
]
```

#### [type](https://runware.ai/docs/models/xai-grok-4-3#request-tools-type)

- **Path**: `tools.type`
- **Type**: `string`
- **Required**: true

Kind of tool to make available to the model. User-defined functions require `name` and `schema`, while built-in tools (`search`, `codeInterpreter`) are executed server-side by the provider.

**Allowed values**: `function`

#### [name](https://runware.ai/docs/models/xai-grok-4-3#request-tools-name)

- **Path**: `tools.name`
- **Type**: `string`
- **Max**: `64`

Unique function name. Required for function tools.

#### [description](https://runware.ai/docs/models/xai-grok-4-3#request-tools-description)

- **Path**: `tools.description`
- **Type**: `string`

Explanation of what the function does, used by the model to decide when to call it.

#### [schema](https://runware.ai/docs/models/xai-grok-4-3#request-tools-schema)

- **Path**: `tools.schema`
- **Type**: `object`

JSON Schema object describing the function's input parameters.

## Response Parameters

### [taskType](https://runware.ai/docs/models/xai-grok-4-3#response-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `textInference`

Identifier for the type of task this response belongs to.

### [taskUUID](https://runware.ai/docs/models/xai-grok-4-3#response-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier echoed from the original request, used to match async responses to their tasks.

### [text](https://runware.ai/docs/models/xai-grok-4-3#response-text)

- **Type**: `string`
- **Required**: true

Generated text content.

### [cost](https://runware.ai/docs/models/xai-grok-4-3#response-cost)

- **Type**: `float`

Task cost in USD. Present when `includeCost` is set to `true` in the request.

### [finishReason](https://runware.ai/docs/models/xai-grok-4-3#response-finishreason)

- **Type**: `string`
- **Required**: true

The reason why the model stopped generating tokens.

**Possible values**: `stop` `length` `content_filter` `unknown`

### [usage](https://runware.ai/docs/models/xai-grok-4-3#response-usage)

- **Path**: `usage.promptTokens`
- **Type**: `object (4 properties)`
- **Required**: true

Token usage statistics for the request.

#### [promptTokens](https://runware.ai/docs/models/xai-grok-4-3#response-usage-prompttokens)

- **Path**: `usage.promptTokens`
- **Type**: `integer`
- **Required**: true
- **Min**: `0`

Number of tokens in the input prompt.

#### [completionTokens](https://runware.ai/docs/models/xai-grok-4-3#response-usage-completiontokens)

- **Path**: `usage.completionTokens`
- **Type**: `integer`
- **Required**: true
- **Min**: `0`

Number of tokens generated in the response.

#### [totalTokens](https://runware.ai/docs/models/xai-grok-4-3#response-usage-totaltokens)

- **Path**: `usage.totalTokens`
- **Type**: `integer`
- **Required**: true
- **Min**: `0`

Total number of tokens used (prompt + completion).

#### [thinkingTokens](https://runware.ai/docs/models/xai-grok-4-3#response-usage-thinkingtokens)

- **Path**: `usage.thinkingTokens`
- **Type**: `integer`
- **Min**: `0`

Number of tokens used for internal reasoning. Billed separately.

## Examples

### text-generation (text-generation)

```
Why is this study being done?

The sponsor is testing an investigational medicine called RZ-417 in adults with moderate-to-severe rheumatoid arthritis whose symptoms continue despite stable methotrexate treatment. RZ-417 is not approved by the U.S. Food and Drug Administration for rheumatoid arthritis or any other condition. About 180 people will take part. The study lasts up to 28 weeks: up to 4 weeks to check if you can join, 20 weeks of study treatment, and one safety visit about 4 weeks after the last dose.

What will happen if I join?

If you join, you will continue your current stable dose of methotrexate. You will be assigned by chance to get either RZ-417 or placebo. A placebo looks like the study medicine but contains no active RZ-417. Two out of every three people will get RZ-417 and one out of every three will get placebo. Neither you nor the study team will know which one you receive while the study is running, unless the information is needed right away for an urgent medical problem.

You will come to the clinic 11 times. At these visits the team will review your medical history, do physical and joint exams, check vital signs, ask questions about pain, fatigue, daily activities, and quality of life, collect blood and urine, do pregnancy testing if needed, and record an electrocardiogram at the first and last treatment visits. About 240 mL of blood will be taken over the whole study. Some of the blood tests measure study-drug levels and immune responses; these tests are for research only and would not be done in usual arthritis care. You will also record symptoms, medicines, and any injection reactions in an electronic diary. Each visit should take 1 to 3 hours. You will get study injections under the skin every 2 weeks.

What are the risks and discomforts?

In earlier studies the most common reactions linked to RZ-417 were injection-site pain, redness, swelling, headache, nausea, and upper respiratory infections. Because RZ-417 affects the immune system, it may raise the chance of infection, including serious infection. Allergic reactions, including rare severe ones, can occur. Long-term risks and risks that are uncommon or still unknown cannot be ruled out because the medicine is investigational.

Blood draws can cause brief pain, bruising, light-headedness, fainting, or, rarely, infection. Electrocardiogram stickers may irritate the skin. Filling out questionnaires may cause boredom or emotional discomfort. Methotrexate has its own risks; your regular doctor will discuss those with you.

The effects of RZ-417 on an embryo, fetus, or nursing infant are unknown. People who could become pregnant must have a negative pregnancy test before starting and use a highly effective birth-control method during treatment and for 12 weeks after the last dose. People who could father a child must use effective birth control during the same time. Pregnancy must be reported to the study team right away. Pregnant or breastfeeding individuals cannot join.

Could I benefit?

You may or may not feel any improvement in your arthritis. The information learned may help future patients, but no benefit is promised.

What other choices do I have?

You do not have to join. You can continue your current treatment, change to other approved medicines with your clinician, or look for another study if you are eligible.

Can I leave the study?

Joining is voluntary. Saying no will not affect your regular medical care
```

Finish reason: `length`

```json
{
  "taskType": "textInference",
  "taskUUID": "c320b53e-25a7-469d-8bb6-79e9c2cdffe2",
  "model": "xai:grok@4.3",
  "settings": {
    "thinkingLevel": "high",
    "maxTokens": 700
  },
  "messages": [
    {
      "role": "user",
      "content": "Rewrite the source material below as a participant-facing consent section for a Phase 2 rheumatoid arthritis study. This will be reviewed by an institutional review board and used during enrollment at U.S. research sites.\n\nWrite for an eighth-grade reading level while preserving every material fact, number, uncertainty, and distinction between research and standard care. Use respectful, neutral language. Do not imply that the study drug is effective, minimize risks, invent safeguards, or add medical facts not supplied here.\n\nOrganize the finished section under these exact headings:\n\nWhy is this study being done?\nWhat will happen if I join?\nWhat are the risks and discomforts?\nCould I benefit?\nWhat other choices do I have?\nCan I leave the study?\nWill I be paid or charged?\n\nDefine “placebo” at first use. Clearly explain the 2-to-1 random assignment in everyday language, state that neither the participant nor the study team will know the assignment during the study except when medically necessary, and distinguish research-only procedures from usual arthritis care. Use short paragraphs and bullets where they improve comprehension. Address the reader as “you.” Return only the finished consent section.\n\nSOURCE MATERIAL\n\nPurpose and design: The sponsor is evaluating RZ-417, an investigational injectable medicine, in adults with moderate-to-severe rheumatoid arthritis whose symptoms remain active despite stable methotrexate treatment. RZ-417 has not been approved by the U.S. Food and Drug Administration for rheumatoid arthritis or any other condition. Approximately 180 participants will be enrolled. Participation lasts up to 28 weeks: a screening period of up to 4 weeks, 20 weeks of study treatment, and a safety follow-up visit about 4 weeks after the final dose.\n\nAssignment and blinding: Eligible participants will be randomized in a 2:1 ratio to receive either RZ-417 or placebo by injection under the skin once every 2 weeks. Assignment is determined by chance. Participants will continue their prescribed stable dose of methotrexate. The study is double-blind. Participants, investigators, and most study personnel will not know treatment assignment unless the information is needed to manage an urgent medical problem. A placebo contains no active RZ-417.\n\nVisits and procedures: Participants will attend 11 clinic visits. Procedures include medical-history review; physical examinations; vital signs; joint examinations; questionnaires about pain, fatigue, daily activities, and quality of life; blood and urine collection; pregnancy testing when applicable; electrocardiograms at screening and Week 20; and study-drug or placebo injections. Approximately 240 mL of blood will be collected over the entire study. Some blood samples will measure study-drug levels and immune responses; these tests are for research and would not ordinarily be part of routine arthritis care. Participants will record symptoms, medications, and injection reactions in an electronic diary. Each visit is expected to last 1 to 3 hours.\n\nKnown and possible risks: In earlier studies, commonly reported reactions associated with RZ-417 were injection-site pain, redness, swelling, headache, nausea, and upper respiratory infections. Because RZ-417 changes immune-system activity, it may increase susceptibility to infection. Serious infection is possible. Allergic reactions, including rare severe reactions, may occur. Long-term risks and risks that are uncommon or not yet known cannot be ruled out because the medicine is investigational. Blood draws may cause temporary pain, bruising, light-headedness, fainting, or, rarely, infection. Electrocardiogram stickers may irritate the skin. Answering questionnaires may cause boredom or emotional discomfort. Methotrexate has its own known risks, which the treating clinician will discuss separately.\n\nReproductive precautions: The effects of RZ-417 on an embryo, fetus, or nursing infant are unknown. Participants who could become pregnant must have a negative pregnancy test before treatment and use a highly effective contraceptive method during treatment and for 12 weeks after the final dose. Participants who could father a child must use effective contraception during the same period. Pregnancy must be reported to the study team immediately. Pregnant or breastfeeding individuals cannot participate.\n\nBenefits and alternatives: Participants may or may not experience improvement. Information learned may help future patients, but no benefit is promised. Alternatives include not joining, continuing current treatment, changing approved medicines with the participant’s clinician, or joining another study if eligible.\n\nVoluntary participation and withdrawal: Participation is voluntary. Refusal will not affect access to regular medical care or benefits. Participants may leave at any time without penalty. The investigator may end participation for safety concerns, failure to follow study requirements, pregnancy, study cancellation, or other reasons allowed by the protocol. Participants who stop treatment may be asked to complete safety assessments, but they may decline.\n\nCosts and payment: The sponsor will provide RZ-417 or placebo and research-only tests at no charge. Participants or their insurers remain responsible for usual medical care, including methotrexate and treatment unrelated to the research. Participants will receive $75 for each completed clinic visit, up to $825 total. Reasonable parking or public-transit expenses will be reimbursed with receipts, up to $30 per visit."
    }
  ]
}
```

---

### tool-calling (tool-calling)

```
{
  "taskType": "textInference",
  "taskUUID": "78aa0396-d278-44f0-b92d-f9aa681292d8",
  "cost": 0.0002359,
  "finishReason": "tools",
  "reasoningContent": [
    "The task is: \"Open the official action log for recall RC-2026-017 covering Cardiomab 50 mg prefilled syringes, lots CM50-24081 and CM50-24082. Retrieve the current UTC date and time from the authorita..."
  ],
  "tools": [
    {
      "id": "call-dfb29085-8ca9-4646-b6ff-ee982af774a6-0",
      "name": "get_current_utc_time"
    }
  ]
}
```

Finish reason: `tools`

```json
{
  "taskType": "textInference",
  "taskUUID": "78aa0396-d278-44f0-b92d-f9aa681292d8",
  "model": "xai:grok@4.3",
  "settings": {
    "temperature": 0.2,
    "thinkingLevel": "low",
    "maxTokens": 300
  },
  "messages": [
    {
      "role": "user",
      "content": "Open the official action log for recall RC-2026-017 covering Cardiomab 50 mg prefilled syringes, lots CM50-24081 and CM50-24082. Retrieve the current UTC date and time from the authoritative time service and record it as the recall activation timestamp. Do not estimate the time or derive it from the conversation."
    }
  ],
  "tools": [
    {
      "type": "function",
      "name": "get_current_utc_time",
      "description": "Returns the current date and time in UTC.",
      "schema": {}
    }
  ],
  "toolChoice": {
    "type": "tool",
    "name": "get_current_utc_time"
  }
}
```

---

### text-generation (text-generation)

```
**Executive Recommendation**  
Policy C is the only option that stays inside the churn limit and the operational constraints while still producing a plausible path to the 12% ARR target. It requires cohort-specific rules: Starter and Team move at renewal to list pricing (Core at $10/seat with 3-seat minimum, Pro at $12/seat with 10-seat minimum); Business receives a one-year 20% ARR cap at renewal, then moves to Advanced list pricing ($14/seat, 25-seat minimum) on the subsequent renewal. The 120 largest Business accounts receive the first 180 high-touch calls each month; the remaining Business accounts and all Team accounts over 20 seats receive email-plus-self-serve migration. Enterprise contracts stay untouched. Annual prepayment keeps the two-month discount.

**Cohort Economics**  
Average monthly price before discounts and annual adjustment:  
- Starter (3 seats): $39 → $30 (Core), –23%.  
- Team (11 seats): $99 → $132 (Pro), +33%.  
- Business (31 seats): $249 → $434 (Advanced), +74%.  

These are simple arithmetic conversions of the stated average seat counts to the new per-seat rates; they do not incorporate existing discounts, within-cohort seat variance, or renewal timing. Consequently, they cannot be read as expected revenue outcomes or churn predictors. They only illustrate directional pressure on the two cohorts that will see increases.

**Policy Comparison**  
Policy A applies list pricing at the next renewal for all non-Enterprise customers. Business customers face a 74% jump; given the survey result that only 28% view increases above 25% as fair, logo churn is likely to exceed the 0.4-point cap for that cohort. Policy B defers every migration by 12 months, so no new revenue materializes inside the first 12 months and the 12% ARR goal is missed by construction. Policy C limits the first-year Business increase to 20% and therefore keeps the immediate price signal inside the fairness threshold reported in the survey, while still allowing full list pricing one renewal later.

**Migration Design**  
One migration rule per legacy cohort satisfies the engineering constraint. At renewal: Starter customers map to Core list; Team customers map to Pro list; Business customers map to a 20%-capped contract that converts to Advanced list on the next renewal. Annual prepayment continues to receive two months free on the new list price. The 180 Customer Success calls per month are allocated first to the 120 largest Business accounts (52% of Business ARR), then to the remaining Business accounts ordered by ARR, then to Team accounts with above-average seat counts. All other customers receive standardized email and in-product prompts.

**Risk Controls**  
Exception policy: any customer whose capped Business renewal still exceeds 25% receives a one-time additional 5% ARR credit for the first year only; this credit is approved by the Head of CS in batches of no more than 30 per month. Stop/go thresholds: if any cohort’s monthly logo churn rises more than 0.3 points above its trailing three-month average for two consecutive months, pause further migrations and escalate to the pricing committee within five business days. Missing data to collect before final approval: actual discount percentages by cohort, renewal-month distribution over the next 18 months, and seat-count histograms for the top 200 Business accounts.

**90-Day Plan**  
Days 1
```

Finish reason: `length`

```json
{
  "taskType": "textInference",
  "taskUUID": "92746ab4-ff1f-471f-9067-eeae8c92d14a",
  "model": "xai:grok@4.3",
  "settings": {
    "thinkingLevel": "high",
    "maxTokens": 700
  },
  "messages": [
    {
      "role": "user",
      "content": "Act as a senior B2B SaaS pricing strategist. Write a board-ready decision memo for Northstar Workflow, which plans to replace bundled plans with per-seat pricing on January 1. Use only the facts below; do not introduce external benchmarks.\n\nCurrent annualized revenue is $8.0M. Customer mix and performance:\n- Starter: 44% of customers, $1.35M ARR, $39/month for up to 5 seats, average 3 active seats, 2.4% monthly logo churn.\n- Team: 38% of customers, $2.85M ARR, $99/month for up to 20 seats, average 11 active seats, 1.5% monthly logo churn.\n- Business: 14% of customers, $2.40M ARR, $249/month for up to 50 seats, average 31 active seats, 0.9% monthly logo churn.\n- Enterprise: 4% of customers, $1.40M ARR, negotiated contracts. Enterprise pricing will not change.\n\nProposed plans:\n- Core: $10 per active seat per month, 3-seat minimum.\n- Pro: $12 per active seat per month, 10-seat minimum.\n- Advanced: $14 per active seat per month, 25-seat minimum.\n- Annual prepayment receives two months free.\n- Intended mapping is Starter to Core, Team to Pro, and Business to Advanced.\n\nManagement objectives and constraints:\n- Increase total ARR by at least 12% within 12 months.\n- Keep any increase in monthly logo churn below 0.4 percentage points for each cohort.\n- Do not change signed Enterprise contracts.\n- Billing engineering can support only one migration rule per legacy cohort.\n- Customer Success can conduct at most 180 high-touch migration calls per month.\n- The 120 largest Business accounts represent 52% of Business ARR.\n- A recent survey found that 61% of Team and Business respondents accept per-seat pricing in principle, but only 28% consider a first-year increase above 25% fair.\n- No reliable price-elasticity study exists.\n\nEvaluate these three policies:\nA. Move every non-Enterprise customer to list pricing at the next renewal.\nB. Grandfather all existing customers for 12 months, then move them to list pricing.\nC. Move customers at renewal but cap the first-year increase at 20%; apply full list pricing at the following renewal.\n\nShow the implied monthly price change for the average customer in each legacy cohort before discounts. Explain what can and cannot be concluded from those averages. Recommend one policy, including any cohort-specific migration rules that remain operationally feasible. Provide a practical exception policy, a prioritized allocation of the 180 monthly Customer Success calls, customer-communication principles, measurable stop/go thresholds, and a 90-day implementation plan.\n\nWrite in direct natural language with these sections: Executive Recommendation, Cohort Economics, Policy Comparison, Migration Design, Risk Controls, and 90-Day Plan. Distinguish calculations from assumptions, avoid false precision, surface tensions with the 12% ARR objective, and identify any missing data that should be collected before final approval."
    }
  ],
  "toolChoice": {
    "type": "none"
  }
}
```

---

### tool-calling (tool-calling)

```
{
  "taskType": "textInference",
  "taskUUID": "1312b239-3d56-4998-ad89-f9d64ed5165c",
  "cost": 0.0002059,
  "finishReason": "tools",
  "reasoningContent": [
    "The task is: \"Our international desk is lifting the embargo on the Helios Energy acquisition announcement immediately. Retrieve the current UTC date and time in RFC 3339 format, precise to the second,..."
  ],
  "tools": [
    {
      "id": "call-7f1252fe-af96-48a9-a304-de6bf8c69efc-0",
      "name": "get_current_utc_time"
    }
  ]
}
```

Finish reason: `tools`

```json
{
  "taskType": "textInference",
  "taskUUID": "1312b239-3d56-4998-ad89-f9d64ed5165c",
  "model": "xai:grok@4.3",
  "settings": {
    "temperature": 0.2,
    "thinkingLevel": "low",
    "maxTokens": 300
  },
  "messages": [
    {
      "role": "user",
      "content": "Our international desk is lifting the embargo on the Helios Energy acquisition announcement immediately. Retrieve the current UTC date and time in RFC 3339 format, precise to the second, so our publishing system can record the official release timestamp."
    }
  ],
  "tools": [
    {
      "type": "function",
      "name": "get_current_utc_time",
      "description": "Returns the current date and time in UTC.",
      "schema": {}
    }
  ],
  "toolChoice": {
    "type": "tool",
    "name": "get_current_utc_time"
  }
}
```