Skip to content

xAI · Model profile

Grok 4

Intelligence, Performance & Price Analysis

Source: Canonical slug: grok-4 · Canonical model registry at build time

Routing verdict

Frontier research / complex reasoning

Reserve this model for the hardest requests; simpler or repetitive work should stay on cheaper routes.

Grok 4 summary metrics

Intelligence
33 scoreGrok 4 scores 33 (estimated) on the Artificial Analysis Intelligence Index, placing it above average among other reasoning models in a similar price tier (median: 28).
Input Price
$5.50 / 1M tokensGrok 4 costs $5.50 per 1M input tokens (at the higher end, median: $1.50) and $27.50 per 1M output tokens (at the higher end, median: $8.40), based on the median across providers serving the model.
Output Price
$27.50 / 1M tokensGrok 4 costs $5.50 per 1M input tokens (at the higher end, median: $1.50) and $27.50 per 1M output tokens (at the higher end, median: $8.40), based on the median across providers serving the model.

Assessment

Routing verdict and task fit

Grok 4 scores 33 (estimated) on the Artificial Analysis Intelligence Index, placing it above average among other reasoning models in a similar price tier (median: 28). Grok 4 costs $5.50 per 1M input tokens (at the higher end, median: $1.50) and $27.50 per 1M output tokens (at the higher end, median: $8.40), based on the median across providers serving the model.
Task fit assessment
WorkloadRatingNotes
Complex reasoning & agentic workflowsOptimalIntelligence score 33 places this model among top performers. Suitable for multi-step analysis and agentic loops.
Cost-sensitive pipelinesViableOutput pricing at $27.50 is premium. Route high-volume simple tasks to cheaper alternatives.

Evidence

Evidence charts

10 published comparison charts for this profile. Grok 4 is marked with a diamond and bold label; each chart has a table view and lists its source and retrieval date.

AA-Omniscience Index

AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis

20 models in the published order. Grok 4.3 (high): 18.3, row 8 of 20 in the published order.
  1. Claude Fable 5 (with fallback)40.1
  2. Gemini 3.1 Pro Preview32.9
  3. Claude Opus 4.8 (max)27.4
  4. Claude Opus 4.7 (max)26.2
  5. Gemini 3.5 Flash22.7
  6. GPT-5.5 (xhigh)20.1
  7. GPT-5.5 (high)18.4
  8. Grok 4.3 (high) (this model)18.3
  9. GPT-5.5 (medium)17.3
  10. Claude Sonnet 5 (max)15.3
  11. GPT-5.5 (low)14.6
  12. Qwen3.7 Max14.1
  13. Claude Sonnet 4.6 (max)12.4
  14. Grok Build 0.1 06167.4
  15. Kimi K2.66.4
  16. GPT-5.4 (xhigh)5.7
  17. Muse Spark4.1
  18. GLM-5.2 (max)4
  19. Grok 4 (this model)3.8
  20. GLM-5.11.9

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Chart interpretation
Each bar is one model measured the same way, so this model can be read against the others. The chart lists its source and retrieval date.

Artificial Analysis Intelligence Index by Open Weights / Proprietary

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR · Evaluation results measured independently by Artificial Analysis

20 models in the published order. Grok 4 is not marked in this published set of 20 models.
  1. Claude Fable 5 (with fallback)59.9
  2. Claude Opus 4.8 (max)55.7
  3. GPT-5.5 (xhigh)54.8
  4. Claude Opus 4.7 (max)53.5
  5. Claude Sonnet 5 (max)53.4
  6. GPT-5.5 (high)53.1
  7. GPT-5.4 (xhigh)51.4
  8. GLM-5.2 (max)51.1
  9. GPT-5.5 (medium)50.4
  10. Gemini 3.5 Flash50.2
  11. Claude Sonnet 4.6 (max)47.2
  12. Gemini 3.1 Pro Preview46.5
  13. Qwen3.7 Max46
  14. MiniMax-M344.4
  15. DeepSeek V4 Pro (max)44.3
  16. Kimi K2.644.2
  17. GPT-5.5 (low)43.5
  18. Muse Spark43.1
  19. Kimi K2.7 Code41.9
  20. Claude Sonnet 5 (Non-reasoning)41.7

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Benchmark interpretation
Higher values indicate stronger overall capability. Bars show how models compare on composite intelligence metrics.

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR · Evaluation results measured independently by Artificial Analysis

20 models in the published order. Grok 4 is not marked in this published set of 20 models.
  1. Claude Fable 5 (with fallback)59.9
  2. Claude Opus 4.8 (max)55.7
  3. GPT-5.5 (xhigh)54.8
  4. Claude Opus 4.7 (max)53.5
  5. Claude Sonnet 5 (max)53.4
  6. GPT-5.5 (high)53.1
  7. GPT-5.4 (xhigh)51.4
  8. GLM-5.2 (max)51.1
  9. GPT-5.5 (medium)50.4
  10. Gemini 3.5 Flash50.2
  11. Claude Sonnet 4.6 (max)47.2
  12. Gemini 3.1 Pro Preview46.5
  13. Qwen3.7 Max46
  14. MiniMax-M344.4
  15. DeepSeek V4 Pro (max)44.3
  16. Kimi K2.644.2
  17. GPT-5.5 (low)43.5
  18. Muse Spark43.1
  19. Kimi K2.7 Code41.9
  20. Claude Sonnet 5 (Non-reasoning)41.7

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Benchmark interpretation
Higher values indicate stronger overall capability. Bars show how models compare on composite intelligence metrics.

Intelligence

Artificial Analysis Intelligence Index · Higher is better · Evaluation results measured independently by Artificial Analysis

12 models in the published order. Grok 4.3 (high): 37.6, row 10 of 12 in the published order.
  1. Claude Fable 5 (with fallback)59.9
  2. GPT-5.5 (xhigh)54.8
  3. GLM-5.2 (max)51.1
  4. Gemini 3.5 Flash50.2
  5. MiniMax-M344.4
  6. DeepSeek V4 Pro (max)44.3
  7. Kimi K2.644.2
  8. Muse Spark43.1
  9. Nemotron 3 Ultra37.8
  10. Grok 4.3 (high) (this model)37.6
  11. Grok 4 (this model)33.3
  12. gpt-oss-120b (high)23.8

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Benchmark interpretation
Higher values indicate stronger overall capability. Bars show how models compare on composite intelligence metrics.

Speed

Output tokens per second · Higher is better · Evaluation results measured independently by Artificial Analysis

10 models in the published order. Grok 4.3 (high): 164 t/s, row 5 of 10 in the published order.
  1. gpt-oss-120b (high)314 t/s
  2. Nemotron 3 Ultra250 t/s
  3. GLM-5.2 (max)218 t/s
  4. Gemini 3.5 Flash192 t/s
  5. Grok 4.3 (high) (this model)164 t/s
  6. MiniMax-M399 t/s
  7. GPT-5.5 (xhigh)88 t/s
  8. Kimi K2.676 t/s
  9. DeepSeek V4 Pro (max)72 t/s
  10. Claude Fable 5 (with fallback)71 t/s

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Output speed context
Speed should be read together with latency and cost; fast output does not always mean better user-perceived performance.

Context Window

Context window: tokens limit · Higher is better · Evaluation results measured independently by Artificial Analysis

20 models in the published order. Grok 4.3 (high): 1M, row 6 of 20 in the published order.
  1. GPT-5.4 (xhigh)1.05M
  2. Gemini 3.5 Flash1M
  3. Claude Fable 5 (with fallback)1M
  4. GLM-5.2 (max)1M
  5. DeepSeek V4 Pro (max)1M
  6. Grok 4.3 (high) (this model)1M
  7. MiniMax-M31M
  8. Claude Opus 4.8 (max)1M
  9. Claude Opus 4.7 (max)1M
  10. Claude Sonnet 5 (max)1M
  11. Claude Sonnet 4.6 (max)1M
  12. Gemini 3.1 Pro Preview1M
  13. Qwen3.7 Max1M
  14. Claude Sonnet 5 (Non-reasoning)1M
  15. GPT-5.5 (xhigh)922k
  16. GPT-5.5 (high)922k
  17. GPT-5.5 (medium)922k
  18. GPT-5.5 (low)922k
  19. GPT-5.4 mini (xhigh)400k
  20. Nemotron 3 Ultra262.1k

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Context window sizing
Larger context windows enable complex multi-turn tasks but may increase latency and cost for long inputs.

Cost per Intelligence Index Task

Weighted average cost (USD) per Artificial Analysis Intelligence Index task, segmented by token type. Lower is better · Evaluation results measured independently by Artificial Analysis

20 models in the published order. Grok 4.3 (high): 0.0101, row 4 of 20 in the published order.
  1. DeepSeek V4 Pro (max)0.0063
  2. gpt-oss-120b (high)0.00454
  3. MiniMax-M30.0139
  4. Grok 4.3 (high) (this model)0.0101
  5. Kimi K2.7 Code0.0203
  6. GPT-5.5 (low)0.0475
  7. Grok Build 0.1 06160.0203
  8. GLM-5.10.0426
  9. Nemotron 3 Ultra0.0209
  10. Gemini 3.1 Pro Preview0.0329
  11. GPT-5.5 (medium)0.0754
  12. Kimi K2.60.0216
  13. GLM-5.2 (max)0.0268
  14. GPT-5.4 mini (xhigh)0.0215
  15. Gemini 3.5 Flash0.0884
  16. GPT-5.5 (high)0.0943
  17. GPT-5.5 (xhigh)0.107
  18. GPT-5.4 (xhigh)0.0722
  19. Qwen3.7 Max0.0343
  20. Claude Sonnet 4.6 (max)0.104

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Benchmark interpretation
Higher values indicate stronger overall capability. Bars show how models compare on composite intelligence metrics.
Cost pressure
Ethen Gateway should route away from this model when the task is simple, repetitive, or likely to generate many output tokens.

Cost per Task

Weighted average cost (USD) per Intelligence Index task · Lower is better · Evaluation results measured independently by Artificial Analysis

10 models in the published order. Grok 4.3 (high): $0.139, row 4 of 10 in the published order.
  1. DeepSeek V4 Pro (max)$0.045
  2. gpt-oss-120b (high)$0.061
  3. MiniMax-M3$0.125
  4. Grok 4.3 (high) (this model)$0.139
  5. Nemotron 3 Ultra$0.244
  6. Kimi K2.6$0.346
  7. GLM-5.2 (max)$0.370
  8. Gemini 3.5 Flash$0.586
  9. GPT-5.5 (xhigh)$0.863
  10. Claude Fable 5 (with fallback)$2.75

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Cost pressure
Ethen Gateway should route away from this model when the task is simple, repetitive, or likely to generate many output tokens.

Cost to Run Artificial Analysis Intelligence Index

Cost (USD) to run all evaluations in the Artificial Analysis Intelligence Index · Evaluation results measured independently by Artificial Analysis

20 models in the published order. Grok 4.3 (high): 12.5, row 4 of 20 in the published order.
  1. gpt-oss-120b (high)3.5
  2. DeepSeek V4 Pro (max)7.2
  3. MiniMax-M314.5
  4. Grok 4.3 (high) (this model)12.5
  5. GPT-5.5 (low)65.2
  6. Grok Build 0.1 061620
  7. Nemotron 3 Ultra33.5
  8. Kimi K2.7 Code24.3
  9. GLM-5.140.4
  10. Gemini 3.1 Pro Preview49
  11. GLM-5.2 (max)32.9
  12. Kimi K2.625
  13. GPT-5.5 (medium)98.8
  14. Gemini 3.5 Flash114
  15. GPT-5.4 mini (xhigh)21.9
  16. Qwen3.7 Max47.6
  17. GPT-5.5 (high)118.8
  18. GPT-5.4 (xhigh)88.5
  19. GPT-5.5 (xhigh)131.8
  20. Claude Sonnet 4.6 (max)149.3

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Benchmark interpretation
Higher values indicate stronger overall capability. Bars show how models compare on composite intelligence metrics.
Cost pressure
Ethen Gateway should route away from this model when the task is simple, repetitive, or likely to generate many output tokens.

Pricing: Cache Hit, Input, and Output

Price (USD per M Tokens) · Evaluation results measured independently by Artificial Analysis

20 models; segments: Cache Hit, Input, Output. Grok 4.3 (high): $3.95, row 6 of 20 in the published order.
  • Cache Hit
  • Input
  • Output
  1. gpt-oss-120b (high)$0.900
  2. DeepSeek V4 Pro (max)$1.31
  3. MiniMax-M3$1.56
  4. Grok Build 0.1 0616$3.20
  5. Nemotron 3 Ultra$3.60
  6. Grok 4.3 (high) (this model)$3.95
  7. Kimi K2.6$5.11
  8. Kimi K2.7 Code$5.11
  9. GPT-5.4 mini (xhigh)$5.33
  10. GLM-5.2 (max)$6.06
  11. GLM-5.1$6.06
  12. Qwen3.7 Max$10.25
  13. Gemini 3.5 Flash$10.65
  14. Claude Sonnet 5 (max)$12.20
  15. Claude Sonnet 5 (Non-reasoning)$12.20
  16. Gemini 3.1 Pro Preview$14.20
  17. GPT-5.4 (xhigh)$17.75
  18. Claude Sonnet 4.6 (max)$18.30
  19. Claude Opus 4.8 (max)$30.50
  20. Claude Opus 4.7 (max)$30.50

Published benchmark dataset · Retrieved

Method and source
Methodology
Independent test run by Artificial Analysis on dedicated hardware.
Chart interpretation
Each bar is one model measured the same way, so this model can be read against the others. The chart lists its source and retrieval date.

Specifications

Architecture and limits

Grok 4 technical specifications
SpecificationValueSource
Model typeProprietaryfaq
ReasoningYesfaq
Input modalitiesGrok 4 supports text and image input.faq
Output modalitiesGrok 4 supports text output.faq
Context window260k tokensfaq
Open weights / sourceNo, Grok 4 is proprietary. The model weights are not publicly available.faq
ParametersGrok 4 is a proprietary model and xAI has not disclosed the model size or parameter count.faq
API availabilityYes, Grok 4 is available via API through 1 provider.faq

Method

Methodology & Provenance

This page is rendered from the normalized profile and page JSON for Grok 4.

Benchmark values are preserved as normalized; only layout, disclosure ordering, and typography are adjusted for readability.

Questions

Model FAQ and technical disclosures

Grok 4 was released on July 10, 2025.

Grok 4 was created by xAI.

Grok 4 scores 33 (estimated) on the Artificial Analysis Intelligence Index, placing it above average among other reasoning models in a similar price tier (median: 28).

Grok 4 costs $5.50 per 1M input tokens (at the higher end, median: $1.50) and $27.50 per 1M output tokens (at the higher end, median: $8.40), based on the median across providers serving the model.

Grok 4 costs $5.50 per 1M input tokens and $27.50 per 1M output tokens (based on the median across providers serving the model). For a blended rate (7:2:1 cache hit/input/output ratio), this is $7.70 per 1M tokens. Pricing may vary by provider.

Yes, Grok 4 is a reasoning model. It uses extended thinking or chain-of-thought reasoning to work through complex problems before providing an answer.

Grok 4 supports text and image input.

Grok 4 supports text output.

Yes, Grok 4 supports image input and can analyze, describe, and answer questions about images.

Yes, Grok 4 is multimodal. It can process text and image input and generate text output.

Grok 4 has a context window of 260k tokens. This determines how much text and conversation history the model can process in a single request.

No, Grok 4 is proprietary. The model weights are not publicly available.

Grok 4 is a proprietary model and xAI has not disclosed the model size or parameter count.

Grok 4 achieves a score of 33 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.

Yes, Grok 4 is available via API through 1 provider.

Grok 4 is available through 1 API provider.

Models

  • 01xAIOther xAI models in the catalog.
  • 02LeaderboardsCertified metric ranks, when published.