DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///
DARC.DEV
Tool radar
G

Code · Documentation review

Grok Build

A coding-agent surface for the current Grok model family.

Grok 4.7 is documented as available through Grok Build, Cursor and the Grok API. The release emphasizes longer tasks and verification behavior. Choose this surface when you want to evaluate Grok in an agent workflow, and retain the exact model, effort and execution settings in your test record.

Official documentationReviewed 23 Sept 2026
Source-checked Sources & corrections

A good fit for

  • → Evaluating Grok on repository tasks
  • → Long-running code and office workflows
  • → Comparing harness behavior with the API

Know the limits

  • → Published scores do not predict your repository's pass rate.
  • → Fast variants have different pricing.
  • → Tools and availability depend on the product surface.

Your first useful workflow

01

Choose one task

Freeze a repository and define an independently testable outcome.

02

Record the configuration

Save model, effort, harness version and allowed tools.

03

Review the result

Measure correctness, intervention and total run cost.

The practical take

Evaluate the full harness and model combination, not just the release's benchmark table.

Sources & evidence

Capabilities above are summarized from these official documents. Workflow suggestions are editorial guidance, not a claim of measured product superiority.