Grok Code Fast 2

LLM Provider / Model · High-volume agentic coding where latency and cost trump max intelligence

At a glance

Input price	$0.20
Output price	$1.50
Cache price	$0.02
Price tier	Budget
Context window	256K
Max output	32K
Context tier	128K-500K
Speed tier	Fast
Latency	low
Knowledge cutoff	Nov 2025
Modality	Text
Model ID	grok-code-fast-2
Provider	xAI
HumanEval	91%
MMLU	85%
SWE-Bench	65%
Benchmark	High SWE-bench at very low cost
Released	Dec 2025
Hosting	Closed/API
Capabilities	Tool use, Function calling, High-throughput, Code-specialized, Prompt caching

Tool use, Function calling, High-throughput, Code-specialized, Prompt caching

High-volume agentic coding where latency and cost trump max intelligence

Google Antigravity is built around Gemini 3 Pro; Claude and recent GPT-5 models also work, but Grok Code Fast 2 is not in the supported model list.
Kiro's spec mode is tuned for Claude models — Grok Code Fast 2 may work but spec-driven flows are best with Claude.
OpenAI Codex (VS Code) is tied to OpenAI's GPT-5.1 Codex family — Grok Code Fast 2 is not selectable inside this integration.
Gemini CLI is locked to Google Gemini models — your LLM is Grok Code Fast 2.
Claude Code only works with Claude models — your LLM is Grok Code Fast 2.
Devin uses Cognition's proprietary model internally — your selected Grok Code Fast 2 will be ignored.
OpenAI Codex (Cloud) runs GPT-5.1 Codex internally — your selected Grok Code Fast 2 will be ignored.
Jules runs Gemini models internally — your selected Grok Code Fast 2 will be ignored.
The Claude Code SDK is tuned for Claude models — Grok Code Fast 2 may work via the model adapter but loses Claude-specific features (thinking, prompt caching, sub-agents).
Lovable uses a fixed internal model — your selected Grok Code Fast 2 will be ignored.
Replit Agent 3 uses its own model selection internally — your selected Grok Code Fast 2 will be ignored.

Build a full stack around Grok Code Fast 2 — Flowpicker shows compatibility warnings before you commit.