Roboflow
Grok

SpaceXAI: Grok 4.6

Grok 4.6 Overview

Grok 4.6 is a proprietary reasoning model from xAI aimed at long-running agentic workflows, coding, and knowledge work. It accepts text and image input and returns text, with a 500,000 token context window and a knowledge cutoff of February 1, 2026. The model exposes an adjustable reasoning budget with low, medium, high, and xhigh settings, where high is the default, and it supports function calling, structured outputs, web and X search, and code execution as documented tool behaviors. Its visual capability covers interpreting images supplied alongside text prompts, which places it in the visual question answering and document understanding family, and it can also return object detection boxes as text coordinates when prompted.

xAI characterizes Grok 4.6 as the result of an extended post-training run over the Grok 4.5 lineage rather than a new pretrained base. The described recipe combines curated model-generated reasoning and technical data, engineering data, a revised optimizer, regenerated supervised fine-tuning trajectories, and reinforcement learning across agent environments spanning knowledge work, coding, kernel optimization, web development, and computer-aided design. Parameter count and architecture specifics are not disclosed. Independent measurement from Artificial Analysis places the model at 61 on its Intelligence Index, five points above Grok 4.5.

Grok 4.6 Interactive Demo

Model settings

Thinking level

Max output tokens

Default 65,536 · max 65,536

Sign in to adjust thinking and output length per run.

Results appear here. Add an image or pick an example to run Grok 4.6.

Grok 4.6 Details & Performance

Details

Resources

—

Vision Tasks

CaptioningChart Question AnsweringClassificationDocument Question AnsweringImage TaggingMulti-Label ClassificationOCRVision LanguageVisual Question AnsweringObject Detection

Features

Foundation VisionLLMs with Vision CapabilitiesMultimodal Vision

Usage

Past 30 Days

Performance

Avg. Latency

Grok 4.6 Vision Evals

Vision Evals is Roboflow's ground-truth benchmark: every model runs the same real-world samples across six vision tasks, and answers are scored against ground truth.

Evals updated September 5, 2026Pricing updated September 13, 2026

Overall score#25 of 53
69.5%
Avg cost / sample#38 of 53
$0.0069
Avg speed / sample#21 of 53
7.39s
Avg tokens / sample
2.1K

Strengths and weaknesses

Grok 4.6 averages 69.5% across the six Vision Evals tasks, ranking #25 of 53 models overall.

Its weakest relative showing is Object Detection, ranking #49 of 53 at 20.2%.

At $0.0069 per sample it is the 38th cheapest of the 53 benchmarked models, and its average inference time of 7.4s per sample makes it the 21st fastest.

Performance profile

Field medianGrok 4.6

Field medians: Object Detection 53.9%, Counting 56.8%, Identification 84.4%, OCR 88.7%, Data Extraction 84.5%, Reasoning 54.1%.

Results by task

TaskScoreField (0 to 100)RankCost / sampleSpeed
Object Detection
20.2%
#49 of 53$0.00686.47s
Counting
70.3%
#12 of 53$0.00748.69s
Identification
87.5%
#20 of 53$0.00485.33s
OCR
92.0%
#13 of 53$0.00868.50s
Data Extraction
85.6%
#21 of 53$0.00424.51s
Reasoning (low)
61.6%
#18 of 53$0.008710.29s
Reasoning (high)
61.6%
#28 of 39$0.02737.83s
  • Thinking longer changes nothing: the same reasoning score at high effort for 3.1x the cost and 3.7x the latency.

Price vs. performance

Score vs. cost

Overall benchmark score against estimated cost per sample, on a log scale. Upper-left is the sweet spot: high quality at low cost.

52 models on the current benchmark · scores and efficiency pooled across all six tasks at low effort · Grok 4.6 highlighted

Grok 4.6 scores are from a single run per task; a three-run re-run under the current protocol is pending · Methodology

View all Vision Evals →

Grok 4.6 Pricing

Grok 4.6 costs $2.00 per 1M input tokens and $6.00 per 1M output tokens.

Input$2.00 / 1M tokens
Output$6.00 / 1M tokens
Cached input$0.500 / 1M tokens

Pricing updated Sep 13, 2026

Alternatives to Grok 4.6

Other models worth comparing for similar use cases.

OpenAI
GPT-6 Astra
GPT-6 Astra is a proprietary multimodal reasoning model from OpenAI that accepts text and image input and produces text output. It is positioned as the company's flagship system for long-horizon end-to-end work spanning complex reasoning, software engineering, computer use, browsing, research and document creation. The model exposes a graduated reasoning effort control with low, medium, high, xhigh and max settings, and it accepts a change to that setting partway through a conversation rather than only at request time. It launches as a single tier with no smaller mini or nano variants, carries a context window of roughly 1.05 million tokens with a maximum output of 128,000 tokens, and reports a knowledge cutoff of April 30, 2026.OpenAI reports evaluation results across agentic, scientific and security benchmarks, including 96.0% on GPQA Diamond, 64.6% on Terminal-Bench Science, 72.6% on OSWorld 2.0, and a perfect score on ExploitBench, along with near saturation on FrontierMath Tier 4 and ARC-AGI-3. The model supports computer use, structured outputs, streaming, programmatic tool calling, multi-agent orchestration, prompt caching and persisted reasoning, and it keeps earlier context windows searchable so it can recover requirements or tool outputs from previous turns. OpenAI describes Astra as the first of its models to cross the Critical cybersecurity capability threshold under its Preparedness Framework.
Anthropic
Claude Fable 5.1
Claude Fable 5.1 is a proprietary multimodal model from Anthropic in the Mythos-class tier of the Claude 5 family, positioned above Claude Opus for demanding reasoning and long-horizon agentic work. It accepts text and images as input and returns text, with a one million token context window and a maximum output of 128 thousand tokens. Adaptive thinking is always on, and an effort parameter controls how much reasoning the model applies to a given request. Anthropic reports a reliable knowledge and training data cutoff of June 2026. Claude Fable 5.1 and Claude Mythos 5.1 share the same underlying model; the difference between them is the set of safety classifiers applied to dual-use cybersecurity and biology requests.On the vision side, Anthropic documents improvements in reading dense charts, financial filings, and tables nested inside PDF documents, which extends the model toward document understanding, chart question answering, and spreadsheet and slide work. Reported evaluations cover agentic scientific research on Terminal-Bench-Science 0.1, agentic coding on Terminal-Bench 4.0, computer use on OSWorld 2.0, and multidisciplinary reasoning on Humanity's Last Exam. Model weights are not published.
Anthropic
Claude Opus 5
Claude Opus 5 is a large language model with multimodal vision capabilities developed by Anthropic, released on July 24, 2026 as the fourth model in the Claude 5 family. It sits in the Opus tier of Anthropic's lineup, positioned below the Mythos-class Fable 5 and Mythos 5 models, and is framed by Anthropic as the go-to model for most knowledge work and automation tasks. The model approaches Fable 5's capabilities at roughly half the cost, priced at $5 per million input tokens and $25 per million output tokens. It becomes the default model on Claude Max and the strongest model available on Claude Pro. The model ships with a 1 million token context window and an adjustable "effort" parameter that allows users to trade reasoning depth for speed and token savings. Early enterprise customers reported that Opus 5 achieved comparable performance to Opus 4.8's maximum-reasoning mode while generating significantly fewer tokens on average, and demonstrated higher accuracy on financial modeling tasks with fewer tool calls and less time.Claude Opus 5 supports multimodal inputs including images and text, and is designed for agentic workflows, coding, scientific research, and complex enterprise tasks. Anthropic reports the model scores 10.2 percentage points higher than Opus 4.8 on an internal chemistry benchmark, making it the most capable generally available model for scientific research in the Claude lineup. Cyber classifiers on Opus 5 are designed to intervene approximately 85 percent less often than those on Fable 5, with fallback to Opus 4.8 when a classifier triggers. The model does not retain user data for 30 days, unlike Fable 5. It is available across Anthropic's platforms including Claude Code and Claude Cowork, as well as cloud partners.
Qwen
Qwen3.8 Max
Qwen3.8 Max is the flagship tier of Alibaba's Qwen3.8 family, a sparse mixture-of-experts multimodal model with roughly 2.4 trillion total parameters of which about 95 billion activate per token, which keeps serving cost and latency well below what the total parameter count would imply. It builds on the architectural foundation established by Qwen3.5 and accepts text, images, video, and documents as input while producing text output. Reported context handling reaches close to one million tokens, with a maximum generation length of 131,072 tokens, so the model is aimed at long-horizon agentic work such as repository-scale coding, multi-step research, data analysis, and office document workflows.For vision work the model performs image and video understanding, document and chart interpretation, text recognition inside images, and grounded visual question answering, and Alibaba reports gains concentrated in multimodal and agentic evaluation categories rather than general reasoning. Published figures include 86.6 on Terminal-Bench 2.1, 67.7 on SWE-bench Pro, 93.0 on PaperBench, 82.8 on IFBench, and 92.6 on GPQA Diamond. It is the first model in the Max tier of the Qwen line for which the team states weights will be released publicly, alongside a smaller Qwen3.8 27B checkpoint. No training or safety model card has been published.
MoonshotAI
Kimi K3
Kimi K3 is a sparse Mixture-of-Experts large language model developed by Moonshot AI, with 2.8 trillion total parameters and a 1-million-token context window. The model activates 16 out of 896 experts per token using the Stable LatentMoE framework, and is built on two architectural innovations: Kimi Delta Attention (KDA), a hybrid linear attention mechanism that enables up to 6.3x faster decoding in long-context settings, and Attention Residuals (AttnRes), which selectively retrieves representations across model depth and delivers roughly 25% higher training efficiency. Together with refined training and data recipes, these structural advances yield approximately 2.5x better overall scaling efficiency compared to its predecessor Kimi K2. The model applies quantization-aware training from the supervised fine-tuning stage onward, using MXFP4 weights with MXFP8 activations for hardware compatibility. Thinking mode is always enabled at launch, with reasoning effort configurable via the reasoning_effort field.Kimi K3 supports native visual understanding alongside text, accepting image inputs for tasks that combine software engineering and visual reasoning. It targets long-horizon coding, knowledge work, and agentic workflows, and ships in two variants: K3 Max for general chat and agent tasks, and K3 Swarm Max for large-scale parallel processing across many coordinated sub-agents. The model is compatible with the OpenAI SDK via an OpenAI-compatible API. Full model weights are scheduled for release by July 27, 2026 under a Modified MIT license, following the open-weight pattern established by the Kimi K2 model family. A technical report with full architecture, training, and evaluation details is expected to accompany the weights release.
Meta
Muse Spark 1.3
Muse Spark 1.3 is a proprietary multimodal reasoning model from Meta Superintelligence Labs and the fourth Muse Spark release in five months, arriving on September 2, 2026. It takes text, images, video, and document files as input and returns text, and it operates over a context window of 1,048,576 tokens. Meta trains the model for long-horizon agentic work, so it carries accumulated context and prior tool results forward across many turns, reconciles messy or conflicting inputs, and asks for clarification when a task is underspecified. Visual inputs such as screenshots and video clips feed a reasoning loop that runs against a real execution environment rather than a scripted sequence of steps.The model exposes graded reasoning effort settings. An xhigh configuration is generally available at launch, while a max reasoning configuration aimed at harder reasoning and agentic problems arrives after further safety testing. Artificial Analysis measures Muse Spark 1.3 (max) at 62 on its Intelligence Index and the xhigh configuration at 61, with agentic tool-use evaluations driving most of the gain over Muse Spark 1.2; max reaches 52% on Tau3-Bench Banking by spending more turns and reasoning tokens than xhigh. Prior Muse Spark versions emit bounding box coordinates, transcriptions, and structured field extractions from images on Roboflow Vision Evals.

Other SpaceXAI Grok models

Other versions in the same family as Grok 4.6.

GrokGrok 4.5GrokGrok 4

Deploy Grok 4.6 with an API

Grok 4.6 runs as a hosted REST endpoint through Roboflow Workflows. Pick a task, then hand the prompt to your coding agent or copy the code. Deploying the workflow into a free Roboflow workspace replaces the your-workspace and YOUR_API_KEY placeholders with your own.

Connect your agent to Roboflow (once)

Add the Roboflow MCP server

claude mcp add --transport http roboflow https://mcp.roboflow.com/mcp

Run /mcp and authorize Roboflow in your browser when the OAuth flow opens.

Start a new Claude Code session so the MCP loads, then paste the prompt below (it works the same in any agent).

Deploy this workflow to your Roboflow workspace to use it.

Integrate the Roboflow "Grok 4.6" workflow into my app.

- Endpoint: POST https://serverless.roboflow.com/<your-workspace>/workflows/grok-4-6-captioning
- Auth: send my Roboflow API key as `api_key` in the request body, read from the ROBOFLOW_API_KEY env var (never hardcode).
- Body: { "api_key": ..., "inputs": { `image`: { type: "url" | "base64", value }, `model_api_key`: my provider key } }.

With the Roboflow MCP connected, call `workflows_get` on "grok-4-6-captioning" to read the exact input schema and treat it as the source of truth. A live `workflows_run` for this workflow also needs my SpaceXAI key (`model_api_key`) passed as a runtime parameter; if you don't have it yet, skip the test run — it will fail with a server error without the provider key, which is expected and not a problem with your code — and rely on the schema. Validate the real run via the REST call once the keys below are set. Without the MCP, use the contract above.

Before running the app, set up these keys so it does not error at runtime:
- `ROBOFLOW_API_KEY` (sent as `api_key`) from https://app.roboflow.com/settings/api
- `SPACEXAI_API_KEY` (sent as `model_api_key`) from https://console.x.ai — my SpaceXAI key
Create a .gitignore'd .env with these variables, using placeholder values for any I haven't given you. Then pause and tell me directly, in your reply: the full path to the .env file, exactly which keys I need to paste in, and the link to get each one. Wait for me to confirm I've added them before you run anything. Do not run the app until I confirm.

Then add the integration to my codebase: match my project's language, framework, and conventions; read every key from environment variables (never hardcode); add basic error handling; and include a small runnable example. If you can't tell what language my project uses, ask me.
Installpip install inference-sdk

Deploy this workflow to your Roboflow workspace to use it.

# 1. Import the library
from inference_sdk import InferenceHTTPClient

# 2. Connect to your workflow
client = InferenceHTTPClient(
  api_url="https://serverless.roboflow.com",
  api_key="YOUR_API_KEY"
)

# 3. Run your workflow on an image
result = client.run_workflow(
  workspace_name="your-workspace",
  workflow_id="grok-4-6-captioning",
  images={
    "image": "YOUR_IMAGE.jpg"  # Path to your image file
  },
  parameters={
    "model_api_key": "YOUR_SPACEXAI_API_KEY"
  },
  use_cache=True  # cache workflow definition for 15 minutes
)

# 4. Get your results
print(result)

Deploy this workflow to your Roboflow workspace to use it.

const response = await fetch('https://serverless.roboflow.com/your-workspace/workflows/grok-4-6-captioning', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    api_key: 'YOUR_API_KEY',
    inputs: {
      "image": {"type": "url", "value": "IMAGE_URL"},
      "model_api_key": "YOUR_SPACEXAI_API_KEY"
    }
  })
});

const result = await response.json();
console.log(result);

Deploy this workflow to your Roboflow workspace to use it.

curl --location 'https://serverless.roboflow.com/your-workspace/workflows/grok-4-6-captioning' \
--header 'Content-Type: application/json' \
--data '{
  "api_key": "YOUR_API_KEY",
  "inputs": {
    "image": {"type": "url", "value": "IMAGE_URL"},
    "model_api_key": "YOUR_SPACEXAI_API_KEY"
  }
}'

Grok 4.6 License

Proprietary

Grok 4.6 is proprietary: the weights are not distributed, and the Grok 4.6 license is the vendor's commercial terms of service that you accept when you call the API.

Commercial use
Permitted under the vendor terms, typically metered per token or per request, with the vendor usage policy applying to your inputs and outputs.
Modification
Not available. Grok 4.6 weights are closed, so you can configure prompts and use vendor-hosted fine-tuning where it is offered, but you cannot modify the model itself.
Redistribution
Not permitted. You cannot self-host or resell the model; you build on the hosted API instead.

Vendor terms govern data retention, whether your inputs can be trained on, rate limits, and regional availability, and they can change with notice. Review them if you handle regulated or customer data.

Do I need a commercial license for Grok 4.6?

Proprietary terms are set by the vendor rather than negotiated per project, and no open-source obligation attaches to your code. If you would rather deploy a model whose commercial license is included in your plan — on Roboflow Managed Cloud or a Self-Hosted Inference Server — Roboflow's licensing page lists the supported alternatives to Grok 4.6.

Do not hesitate to reach out with questions for your commercial project — our team will help you start solving business problems on the first call. See Roboflow commercial licensing for the models included in each plan.

Talk to sales

This model is proprietary. The author retains all rights, and use of the model is governed by their specific terms of service or license agreement.

Commercial use depends on the terms set by the model author. Most proprietary commercial models require a paid subscription, API key, or per-call billing. Check the provider’s pricing and terms-of-service for details.

License information is provided as a guide and is not legal advice.

Frequently Asked Questions About Grok 4.6 Vision

Yes. Grok 4.6 accepts image input and handles OCR, data extraction, object counting, identification, visual reasoning, and object detection. On Roboflow's Vision Evals its strongest task is Counting at 70.3% (#12 of 53 at low effort). You can test it on your own image in the demo above.

Yes, and it is one of the model's strongest vision skills: its transcriptions match the ground truth 92% on average (#13 of 53 at low effort) on Vision Evals OCR. Pulling specific fields out of documents (data extraction) scores 85.6%.

Not its strength. On Vision Evals, Grok 4.6 scores 20.2% mAP@50 on object detection (#49 of 53 at low effort) and 70.3% judge-graded accuracy on object counting. For production counting or precise localization, pairing it with a specialized detector like RF-DETR or your own trained model in a Roboflow Workflow is usually more reliable: detect the objects, then count the detections.

On our benchmark's task mix, Grok 4.6 averages $0.0069 per sample at $2.00 per 1M input and $6.00 per 1M output tokens (#38 of 53 on cost), with an average speed of 7.4s per sample across the benchmark. Actual cost depends on your images and prompts.

On the overall Vision Evals ranking, Grok 4.6 sits #25 of 53 at 69.5%, just behind Gemini 3.5 Flash-Lite (70.3%) and just ahead of Qwen3.5 35B-A3B (69.4%). See the full side-by-side: Grok 4.6 vs Gemini 3.5 Flash-Lite.