ClosedxAI/10 trustVerified

Grok Code Fast 1

Name: Grok Code Fast 1
Rating: 7

A fast, low-cost reasoning model built for agentic coding workflows.

Visit official page →

About

Grok Code Fast 1 is xAI's first model purpose-built for agentic coding, trained from scratch on a programming-rich corpus with post-training on real-world pull requests and coding tasks. Its strength is speed: xAI optimized serving and tool-call loops so it can iterate through grep, terminal, and file-edit calls with minimal latency. The tradeoff is deliberate - xAI positions it as a daily-driver coder for everyday tasks rather than the strongest reasoner, and early coverage noted it trades some accuracy for velocity. It is strong in TypeScript, Python, Java, Rust, C++, and Go, and was tuned inside real coding platforms via launch partners.

Specs

Context window

256K tokens

Pricing

$0.20 per million input tokens, $1.50 per million output tokens, $0.02 per million cached input tokens (published by xAI, Aug 2025)

Free quota

Free for a limited time via launch partners (GitHub Copilot, Cursor, Cline, Roo Code, Kilo Code, opencode, Windsurf); none on the API itself

API access

Generally available via the xAI API (OpenAI-compatible). Free limited-time access through launch partners: GitHub Copilot, Cursor, Cline, Roo Code, Kilo Code, opencode, and Windsurf. No web search or image input in the original release; a multimodal variant with parallel tool calling and extended context was reported as in training.

Weights

Closed — API only

Released

August 28, 2025

Benchmarks

SWE-Bench-Verified (full subset, xAI's internal harness): 70.8%. Speed (provider-measured, final response tokens only): ~160 tokens/sec, claimed faster than Gemini 2.5 Pro (92.39 TPS) and Qwen3-Coder (80 TPS). xAI pairs public benchmarks with human developer ratings, noting benchmarks do not fully capture agentic workflow quality.

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

codeagentsautomation

Trust score — 7/10

/10 trustVerified

Scored 7/10 (verified tier, metrics-v1). Evidence: Read xAI's full official announcement and model page; SWE-bench 70.8% from xAI's own harness, speed numbers provider-measured; Widely integrated via launch partners at release (Copilot, Cursor, Windsurf); Superseded by Grok Code Fast 1.1 / Grok 4 Fast releases since. License: Proprietary; closed weights, API access only..

How the trust score is built

Full methodology →

Adoption · stars/downloads, log-scaled vs peers: 7.0/10
Maintenance · recency, releases, issue load: 8.0/10
Openness · license + docs: 2.0/10
Manual review · sanity read + red-flag check: 9.0/10

· Read xAI's full official announcement and model page
· SWE-bench 70.8% from xAI's own harness, speed numbers provider-measured
· Widely integrated via launch partners at release (Copilot, Cursor, Windsurf)
· Superseded by Grok Code Fast 1.1 / Grok 4 Fast releases since