Grok 4.3 and Gemini 3.8 Flash are both 1M-context strong-tier models, and neither is cheaper on every token: Gemini wins on input, Grok on output. Gemini's current rate is also an introductory price that ends on December 31, 2026.
Until December 31, input-heavy work is cheaper on Gemini 3.8 Flash and output-heavy work on Grok 4.3. The break-even is about 2.5 input tokens per output token. From January 1, Gemini 3.8 Flash reverts to $1.50/$7.50 per 1M, which makes Grok 4.3 ($1.25/$2.50) cheaper on both.
| Grok 4.3 | Gemini 3.8 Flash | |
|---|---|---|
| Provider | SpaceXAI | |
| Tier | Mid-tier | Mid-tier |
| Input / 1M tokens | $1.25 | $0.75 |
| Output / 1M tokens | $2.50 | $3.75 |
| Cache hit / 1M | $0.20 | $0.075 |
| Cache write / 1M | — | — |
| Context window | 1M | 1M |
| Notes | Reasoning workhorse; replaced Grok 4; ≥200K tokens: $2.50/$5, cache $0.40 | On the Gemini API free and paid tiers; Gemini 3.7 Flash retired Oct 8, 2026 and now auto-routes here (requests still billed at this rate); intro price thru Dec 31, reverts to $1.50/$7.50 (cache $0.15) Jan 1 |
Gemini 3.8 Flash is 40% cheaper on input tokens ($0.75 vs $1.25 per 1M). Grok 4.3 is 33% cheaper on output tokens ($2.50 vs $3.75 per 1M).
| Workload | Grok 4.3 | Gemini 3.8 Flash | Difference |
|---|---|---|---|
| Support chatbot 10,000 conversations × (2,000 input + 400 output tokens) | $35.00 | $30.00 | Gemini 3.8 Flash saves $5.00 |
| Document Q&A 1,000 questions × (30,000 input + 800 output tokens) | $39.50 | $25.50 | Gemini 3.8 Flash saves $14.00 |
| Coding agent session 4M input tokens (90% cache hits) + 150,000 output tokens | $1.59 | $1.13 | Gemini 3.8 Flash saves $0.46 |
Standard rates per 1M tokens from data on this site. The coding-agent row assumes 90% of input is read from the prompt cache (full input price where a model has no cache rate). Cache-write fees, batch discounts and long-context surcharges listed in the notes are not included. Updated November 2, 2026.
your workload generates a lot of text relative to what it reads, you are budgeting past January 1, or you want a model whose listed rate is not a time-limited promotion.
you mostly read large inputs and return short answers — document Q&A, extraction, classification — and can re-evaluate before the intro price ends, or you want to prototype on Google's free tier first.
Gemini saves $1.25 minus $0.75 on every 1M input tokens, and Grok saves the difference between $3.75 and $2.50 on every 1M output tokens. The two balance at roughly 2.5 input tokens per output token. A support bot at 5:1 favours Gemini; an agent that writes long code or reports at 1:1 favours Grok. Check your own logs for the real ratio before choosing.
Grok 4.3 bills prompts of 200K tokens or more at $2.50/$5 per 1M, and xAI's model page says the higher rate applies to every token in that request. Google lists no long-context tier for 3.8 Flash, so for 300K–1M-token prompts Gemini's input advantage grows. Cache hits are $0.075 on Gemini versus $0.20 on Grok, another point for Gemini on repetitive input.
Gemini 3.8 Flash's price is introductory through Dec 31, 2026 and then doubles to $1.50/$7.50; Google's pricing page shows the cache rate doubling too. Gemini 3.7 Flash was retired on Oct 8, 2026 and its requests now route to 3.8 Flash at this rate, so anyone who never migrated is already on it. If you sign a yearly budget now, model both prices.
Today Gemini 3.8 Flash is cheaper on input ($0.75 vs $1.25) and Grok 4.3 on output ($2.50 vs $3.75). After Gemini's intro price ends on Dec 31, 2026, Grok 4.3 is cheaper on both.
Yes. Google's pricing page lists free-tier access for Gemini 3.8 Flash, with paid rates applying once you move to the paid tier.
Yes. Prompts of 200K tokens or more are billed at $2.50 input and $5 output per 1M for the whole request.
How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.