| Model | Score | Status | Released | Price in/out | Context | |
|---|---|---|---|---|---|---|
| Grok 4.7 frontier-class coding and knowledge-work agents at mid-tier prices | 8.2 | GA | 2026-09-21 | $2 / $6 | 500K | Review → |
| Grok Build 0.1 cheap plan-mode coding agent with parallel sub-agents | 6.7 | GA | 2026-05-20 | $1 / $2 | 256K | Review → |
| Grok 4.3 real-time research and agentic work at frontier-class value | 7.7 | GA | 2026-04-30 | $1.25 / $2.50 | 1M | Review → |
| Grok 4.20 Multi-Agent single-call parallel deep research with live X data | 7.0 | beta | 2026-03-31 | $1.25 / $2.50 | 1M | Review → |
| Grok 4.20 runtime reasoning toggle with long context and live X data | 6.9 | GA | 2026-03-10 | $1.25 / $2.50 | 1M | Review → |
Grok 4.7 is xAI's flagship model for coding, agentic tasks and knowledge work, announced and released to the public API on 2026-09-21 (xAI's pages now carry the SpaceXAI name). It is a reasoning-always-on model with a new, larger base model than Grok 4.6, a 500,000-token context window, text and image input, and xAI's server-side web and X search tools. xAI's own launch table puts it level with GPT-5.6 Sol on hard coding benchmarks at half Sol's input price and under a third of its output price. The sentence a buyer needs: Grok 4.7 is the most credible coding and knowledge-work model xAI has shipped, priced at $2 / $6 per 1M tokens, but on release day every performance figure is xAI's own, and it gives up Grok 4.3's 1M context and batch discount.
Full review →