Grok 4.7 doubles per-task cost to $3.74 despite unchanged token pricing
Tool · Dev.to · stat: 2.5x xAI's Grok 4.7 doubles the average cost per task to $3.74 despite keeping its predecessor's $2 per million input and $6 per million output token pricing. Testing by…
Tool · Dev.to · stat: 2.5x
xAI's Grok 4.7 doubles the average cost per task to $3.74 despite keeping its predecessor's $2 per million input and $6 per million output token pricing. Testing by Artificial Analysis reveals the new model is highly verbose, generating 240 million tokens compared to Grok 4.6's 94 million to complete the same benchmark index.
Verbose reasoning models make standard API pricing tables obsolete Founders must budget LLM integrations by benchmarking actual task execution rather than relying on raw per-token API pricing.
Artificial Analysis via Dev.to
Per Artificial Analysis benchmarks reported on Dev.to.
Grok 4.7's high verbosity drives up task costs even though xAI kept token rates identical to the previous version.
Every claim ties to a primary source. See our methodology.