flagship

Grok 4 vs Gemini 3 Pro Preview

Side-by-side pricing, context window, features, and per-use-case verdict. Data sourced from LiteLLM, updated regularly.

Specifications

Feature Grok 4 Gemini 3 Pro Preview
Provider xai vertex ai-language-models
Model ID xai/grok-4 gemini-3-pro-preview
Input / 1M tokens $3.00 $2.00 (33% cheaper)
Output / 1M tokens $15.00 $12.00 (20% cheaper)
Cached input / 1M $0.20
Context window 256kk tokens 1049kk tokens
Max output 256kk tokens 66kk tokens
Vision
Function calling
Prompt caching
Web search
JSON schema

Monthly cost comparison

No cache discount applied — pure token costs at three usage scales.

Scale Grok 4 Gemini 3 Pro Preview Winner
Small
1M in / 0.2M out
$6.00 $4.40 Gemini 3 Pro Preview saves 27%
Medium
10M in / 2M out
$60.00 $44.00 Gemini 3 Pro Preview saves 27%
Large
100M in / 20M out
$600 $440 Gemini 3 Pro Preview saves 27%

Category winners

Cheapest input
Gemini 3 Pro Preview
$2.00/M · 33% less than the other
Cheapest output
Gemini 3 Pro Preview
$12.00/M · 20% less than the other
Larger context
Gemini 3 Pro Preview
1049kk tokens · 310% larger window
More features
Gemini 3 Pro Preview
5 / 5 capabilities

Better for…

High-volume RAG & repeated context
$0.20/M cached input — prompt caching makes repeated context retrieval significantly cheaper
Gemini 3 Pro Preview
Short interactions & high-throughput APIs
$12.00/M output — 20% cheaper on output at scale
Gemini 3 Pro Preview
Long-document processing & large codebases
1049k context window vs 256k — 310% larger context
Gemini 3 Pro Preview
Agentic workflows & tool use
Supports 5 of 5 tracked capabilities (vision, function calling, caching, web search, JSON schema)
Gemini 3 Pro Preview

Deep-dive on each model

Pricing data sourced from LiteLLM and refreshed regularly. Always verify with each provider's official pricing page before making business decisions.