xAI logo

Grok 4.20 (Non-Reasoning)

xai/grok-4.20-0309-non-reasoning

Grok model for agentic tool use, reasoning, coding, and live assistance

Source-linkedgrokClosed weightsReleased 2026-03-09
Report data
Context window1M
Maximum output30K
Provider offers4
Knowledge cutoffUndisclosed
Capabilities

What this model supports

ReasoningNo
Tool callingYes
Structured outputYes
AttachmentsYes
Temperature controlYes
Open weightsNo
Creator
xAI
Family
grok
Modalities
Image, Pdf, Text
License
Not documented
Release date
2026-03-09
Model ID
xai/grok-4.20-0309-non-reasoning
Market availability

Provider pricing

Report price →
Providers offering Grok 4.20 (Non-Reasoning)
ProviderProvider model IDContextInput / 1MOutput / 1MCached / 1MFeaturesAll pricing
LLM Gatewaygrok-4-20-non-reasoning2M$1.250$2.500$0.200· T S
View
{
  "cache_read": 0.2,
  "context_over_200k": {
    "cache_read": 0.4,
    "input": 2.5,
    "output": 5
  },
  "input": 1.25,
  "output": 2.5,
  "tiers": [
    {
      "cache_read": 0.4,
      "input": 2.5,
      "output": 5,
      "tier": {
        "size": 200000,
        "type": "context"
      }
    }
  ]
}
xAIgrok-4.20-0309-non-reasoning1M$1.250$2.500$0.200· T S
View
{
  "cache_read": 0.2,
  "context_over_200k": {
    "cache_read": 0.4,
    "input": 2.5,
    "output": 5
  },
  "input": 1.25,
  "output": 2.5,
  "tiers": [
    {
      "cache_read": 0.4,
      "input": 2.5,
      "output": 5,
      "tier": {
        "size": 200000,
        "type": "context"
      }
    }
  ]
}
Azuregrok-4-20-non-reasoning262K$2.000$6.000· T ·
View
{
  "input": 2,
  "output": 6
}
LLM Gatewaygrok-4-20-beta-0309-non-reasoning2M$2.000$6.000$0.200· T S
View
{
  "cache_read": 0.2,
  "context_over_200k": {
    "cache_read": 0.4,
    "input": 2.5,
    "output": 5
  },
  "input": 2,
  "output": 6,
  "tiers": [
    {
      "cache_read": 0.4,
      "input": 2.5,
      "output": 5,
      "tier": {
        "size": 200000,
        "type": "context"
      }
    }
  ]
}
Evidence

Published benchmark results

Insufficient benchmark coverage. ModelBench does not infer a score from price, context size, or model name.