模型广场

Reasoning-oriented Grok variant built for multi-step problem solving and agentic tool use, with strict prompt adherence and low hallucination.

上下文

2M

输入 / 1M tokens

$2.5

输出 / 1M tokens

$5

缓存读 / 1M

$0.4

缓存写 / 1M

$0

Multi-agent variant that dispatches sub-agents in parallel for deep research, multi-step reasoning, and tool orchestration.

上下文

2M

输入 / 1M tokens

$4

输出 / 1M tokens

$12

缓存读 / 1M

$0.4

缓存写 / 1M

$0

Reasoning-oriented Grok variant built for multi-step problem solving and agentic tool use, with strict prompt adherence and low hallucination.

上下文

2M

输入 / 1M tokens

$2.5

输出 / 1M tokens

$5

缓存读 / 1M

$0.4

缓存写 / 1M

$0

Grok 4.3 is a reasoning model from xAI that supports text and image inputs with text output. It is suited for agentic workflows, instruction following, long-document analysis, and deep research. The model supports a 1M token context window. Requests above 200K total tokens are billed at a higher rate.

上下文

1M

输入 / 1M tokens

$2.5

输出 / 1M tokens

$5

缓存读 / 1M

$0.4

缓存写 / 1M

$0

Grok's flagship reasoning model, running an internal chain of thought before responding. Strong performance on coding and STEM tasks, with support for text, image, and file inputs. Suited for technical problem-solving and long-document analysis.

上下文

500K

输入 / 1M tokens

$2

输出 / 1M tokens

$6

缓存读 / 1M

$0.5

缓存写 / 1M

$0

Grok's agent-oriented release, built for long-running multi-step work across a codebase. Accepts text, image, and file inputs. Suited for research and interactive or visual output.

上下文

500K

输入 / 1M tokens

$2

输出 / 1M tokens

$6

缓存读 / 1M

$0.5

缓存写 / 1M

$0