xAI's Lineup in 2026: Just Grok 4
TokenRate's new Compare Prices grid puts every model's per-token rates, context window, and quality score in a single side-by-side view. xAI's hosted lineup in 2026 is short — just Grok 4 — so the more useful comparison isn't within-provider but across-provider at the same price point. Grok 4 sits at $3.00 input / $15.00 output per 1M tokens, a 256K context, and a blended quality score of 79. That price point puts Grok 4 in direct competition with Claude Sonnet 4.7 and just below Claude Opus 4 — the comparison this guide walks through. Related reading: filter LLM models by tier, cost, quality, Value column vs tokens-per-dollar, and how to pick an LLM by quality score and cost.
Grok 4 in the Compare Prices Grid
Open /tools/compare-prices, click the **xAI** dropdown, and check **Grok 4**. The row shows: input $3.00/1M, output $15.00/1M, 256K context, quality 79, tier **flagship**. The Value column (quality ÷ input cost) reads 26.3 — competitive but not class-leading at the price point. The 256K context is the longest among models priced at $3/1M input, which is one of Grok 4's distinguishing features. The output-to-input ratio is 5×, identical to Claude Sonnet 4.7 — meaning generation-heavy workloads cost the same multiple of input as on Claude.
Cross-Provider Competitors at Grok 4's Price Point
Add the cross-provider competitors to the same grid: **Claude Sonnet 4.7** ($3.00 input / $15.00 output, 200K context, Q80, value 26.7) and **Grok 4** at the same input rate. Above the tier: **Claude Opus 4** ($15.00 input / $75.00 output, 200K context, Q85, value 5.7). Below the tier: **Gemini 2.5 Pro** ($1.25 input / $10.00 output, 1M context, Q78, value 62.4) and **GPT-5** ($1.25 input / $10.00 output, 200K context, Q82, value 65.6). Grok 4 occupies a narrow band: same price as Sonnet 4.7 but 1 quality point below, 0.6× the price of Opus 4 with a 6-point quality gap, 2.4× the price of GPT-5 with a 3-point quality gap. The grid surfaces the squeeze visually.
When Grok 4 Wins in the Grid
Grok 4 wins when: (1) context length matters — its 256K window beats Sonnet's 200K and GPT-5's 200K, useful for codebase QA and document libraries that don't quite fit the 200K bracket but don't need Gemini's 1M; (2) the workload has a real-time-search or X/Twitter-data flavor that xAI specializes in; (3) you want a third opinion alongside Anthropic and OpenAI for routing diversification. Where Grok 4 loses: pure quality-per-dollar (GPT-5 dominates at $1.25 input); long-context-per-dollar (Gemini 2.5 Pro at 1M context for half the price); peak quality (Opus 4 at Q85).
Operationalizing the Grok 4 Pick
Once Grok 4 is in your grid alongside its cross-provider peers, run your projected token volume through /tools/api-cost-estimator to model the monthly bill. For most teams, Grok 4 ends up being a routing diversification pick rather than a primary — 10-20% of traffic to Grok 4 as a hedge against single-vendor risk on Anthropic or OpenAI, with the majority on whichever wins the Value column for the workload. The grid pulls prices live from OpenRouter and quality from a blended Arena AI + Artificial Analysis pipeline — both refresh on a 60-minute incremental cache. Open /tools/compare-prices now, pick your provider dropdowns, and pin the shortlist that matches your workload.