Skip to main content
Source: xAI model docs. Search includes Web Search and X Search, priced at $2.50-$5/1K calls. Cached input: $0.05/1M.

​
Grok 4

grok-4

Reasoning · Speedxai/grok-4xAI’s flagship reasoning model designed for advanced math, logic, code, and scientific thinking, trained with reinforcement learning to augment reasoning with tools like code interpreter and web browsing.
  • $3 / $15
  • 256K context
  • ~30K max output
  • Text, Image input
  • Always-on thinking
  • Web search
  • Knowledge cutoff Nov 2024

grok-4-fast-reasoning

Reasoning · Speedxai/grok-4-fast-reasoningA cost-efficient reasoning variant of Grok 4 that achieves comparable performance while using 40% fewer thinking tokens, with a 2M token context window.
  • $0.20 / $0.50
  • 2M context
  • ~30K max output
  • Text, Image input
  • Thinking
  • Web search
  • Knowledge cutoff Jun 2025

grok-4-fast-non-reasoning

Reasoning · Speedxai/grok-4-fast-non-reasoningThe non-reasoning variant of Grok 4 Fast for tasks that don’t require deep chain-of-thought, offering frontier-level performance with exceptional token efficiency.
  • $0.20 / $0.50
  • 2M context
  • ~30K max output
  • Text, Image input
  • Web search
  • Knowledge cutoff Jun 2025

​
Grok 4.1 Fast

grok-4-1-fast-reasoning

Reasoning · Speedxai/grok-4-1-fast-reasoningReasoning variant with MCP and Agent Tools API support for building agentic applications.
  • $0.20 / $0.50
  • 2M context
  • ~30K max output
  • Text, Image input
  • Thinking
  • Web search
  • Knowledge cutoff Nov 2024

grok-4-1-fast-non-reasoning

Reasoning · Speedxai/grok-4-1-fast-non-reasoningNon-reasoning variant with MCP and Agent Tools API support for high-throughput agentic workloads.
  • $0.20 / $0.50
  • 2M context
  • ~30K max output
  • Text, Image input
  • Web search
  • Knowledge cutoff Nov 2024