Gemini API Pricing vs Claude API: Per-Task Costs
Gemini API pricing vs Claude API pricing on November 22, 2026: per-token rates, free tiers, caching and batch discounts, and three same-task cost examples.
Josh Pocock7 min read
AI Coding
Gemini vs Claude
Stride AI Academy
Gemini API pricing on November 22, 2026 runs from $0.30 per million input tokens for Gemini 3.5 Flash-Lite to $2 for Gemini 4 Carbon, which rises to $4 on December 1, 2026. Claude's API runs from $0.10 for Haiku 5.5 to $8 for Fable 5.2, with Opus 5.5 at $4. Gemini has a free tier on several models. Claude gives new users free credits instead.
This guide puts both price lists side by side, compares free tiers, caching, batch and search pricing, then runs three identical tasks through both APIs with the arithmetic shown. Everything here was checked on November 22, 2026 against Google's Gemini API pricing page (ai.google.dev/gemini-api/docs/pricing now redirects there) and Anthropic's API pricing docs and models overview. All prices are in US dollars per million tokens (MTok) on each provider's standard paid tier.
Gemini API pricing by model
From Google's pricing page, standard paid tier:
| Model | Model ID | Input | Output (incl. thinking) | Free tier |
|---|---|---|---|---|
| Gemini 4 Carbon | gemini-4-carbon |
$2.00 through Nov 30, 2026; $4.00 from Dec 1 | $10.00; $20.00 from Dec 1 | No |
| Gemini 4 Argon | gemini-4-argon |
$2.00 through Nov 30, 2026; $4.00 from Dec 1 | $10.00; $20.00 from Dec 1 | No |
| Gemini 4 Neon | gemini-4-neon |
$3.00 | $15.00 | No |
| Gemini 4 Hydrogen | gemini-4-hydrogen |
$1.50 | $7.50 | Yes |
| Gemini 3.8 Flash | gemini-3.8-flash |
$0.75 through Dec 31, 2026; $1.50 from Jan 1, 2027 | $3.75; $7.50 from Jan 1 | Yes |
| Gemini 3.5 Flash-Lite | gemini-3.5-flash-lite |
$0.30 | $2.50 | Yes |
Google describes Gemini 4 Carbon as its most capable reasoning model, and Argon as a faster, more efficient Gemini 4 model. It calls 3.8 Flash its most intelligent Flash model, and 3.5 Flash-Lite a cost-efficient model for high-volume tasks. None of these six has a separate price for long prompts.
Two price changes are close. Carbon and Argon double on December 1, 2026. Gemini 3.8 Flash doubles on January 1, 2027. The page also says requests to gemini-2.5-pro and gemini-2.5-flash will route to 3.8 Flash and 3.5 Flash-Lite from December 1, 2026.
Claude API pricing by model
From Anthropic's pricing docs:
| Model | Model ID | Input | Output |
|---|---|---|---|
| Claude Fable 5.2 | claude-fable-5-2 |
$8 | $40 |
| Claude Opus 5.5 | claude-opus-5-5 |
$4 | $20 |
| Claude Sonnet 5.5 | claude-sonnet-5-5 |
$2 | $10 |
| Claude Haiku 5.5 (prompts up to 100,000 tokens) | claude-haiku-5-5 |
$0.10 | $0.50 |
| Claude Haiku 5.5 (prompts over 100,000 tokens) | claude-haiku-5-5 |
$0.50 | $2.50 |
All four have a 1M-token context window. Anthropic's docs suggest starting with Opus 5.5 for most work. Our Claude API pricing guide covers every Claude rate in detail.
Gemini vs Claude API cost: rates side by side
Matching each tier by its role:
| Role | Gemini model | Gemini in/out | Claude model | Claude in/out |
|---|---|---|---|---|
| Top reasoning | Gemini 4 Carbon | $2/$10 now, $4/$20 from Dec 1 | Opus 5.5 | $4/$20 |
| Mid, fast | Gemini 3.8 Flash | $0.75/$3.75 now, $1.50/$7.50 from Jan 1 | Sonnet 5.5 | $2/$10 |
| High volume | Gemini 3.5 Flash-Lite | $0.30/$2.50 | Haiku 5.5 | $0.10/$0.50 |
From December 1, Gemini 4 Carbon and Claude Opus 5.5 have the same per-token price. At the low end, Haiku 5.5 is cheaper per token than Gemini 3.5 Flash-Lite for prompts up to 100,000 tokens.
Free tiers
Gemini: Gemini 4 Hydrogen, 3.8 Flash and 3.5 Flash-Lite are free of charge on the free tier. Gemini 4 Carbon, Argon and Neon aren't available on it. Google's page says free-tier content is used to improve its products, and paid-tier content isn't.
Claude: there's no ongoing free API tier. Anthropic's docs say new users get free credits to test the API, then you pay per token.
Caching and batch discounts
| Gemini | Claude | |
|---|---|---|
| Cache reads | Gemini 4 Carbon $0.10 (doubling Dec 1), 3.8 Flash $0.075 (doubling Jan 1), 3.5 Flash-Lite $0.03 | Opus 5.5 $0.20, Sonnet 5.5 $0.10, Haiku 5.5 $0.01 |
| Cache storage | Hourly storage price on Flash models, such as $0.50 per MTok per hour on 3.8 Flash and $1.00 on 3.5 Flash-Lite | No hourly storage price listed; cache writes cost 1.25x input for 5 minutes, 2x for 1 hour |
| Explicit caching | Gemini 4 models: implicit caching only | Yes, with cache_control |
| Batch | 50% off | 50% off input and output |
The practical difference: Claude lets you mark exactly what to cache, while Google's page says the Gemini 4 models support implicit caching only. That makes Claude's cached costs easier to predict.
Search and grounding
- Gemini: grounding with Google Search gives 5,000 free requests a month, shared across Gemini 3 and newer models, then $14 per 1,000 requests.
- Claude: web search costs $10 per 1,000 searches plus tokens. Web fetch has no extra charge.
Google's page also labels Gemini output prices as including thinking tokens, so a model that thinks longer costs more output.
Same-task cost examples
These examples use identical token counts on both APIs, to compare prices. Real counts will differ, because each provider's tokenizer splits text differently. To turn tokens into dollars, divide by 1,000,000 and multiply by the price per MTok.
Task 1: a coding agent run
2,000,000 input tokens and 60,000 output tokens, no caching.
- Claude Opus 5.5: (2 x $4) + (0.06 x $20) = $8.00 + $1.20 = $9.20
- Gemini 4 Carbon, through Nov 30: (2 x $2) + (0.06 x $10) = $4.00 + $0.60 = $4.60
- Gemini 4 Carbon, from Dec 1: (2 x $4) + (0.06 x $20) = $8.00 + $1.20 = $9.20
- Claude Fable 5.2: (2 x $8) + (0.06 x $40) = $16.00 + $2.40 = $18.40
Caching changes this a lot. In our Claude API pricing guide, the same Opus 5.5 run with explicit caching comes to $2.51.
Task 2: a support chatbot
10,000 messages a month, each with 3,000 input tokens and 400 output tokens, no caching. That's 30 MTok of input and 4 MTok of output.
- Claude Sonnet 5.5: (30 x $2) + (4 x $10) = $60.00 + $40.00 = $100.00 a month
- Gemini 3.8 Flash, through Dec 31: (30 x $0.75) + (4 x $3.75) = $22.50 + $15.00 = $37.50 a month
- Gemini 3.8 Flash, from Jan 1, 2027: (30 x $1.50) + (4 x $7.50) = $45.00 + $30.00 = $75.00 a month
- Claude Haiku 5.5: (30 x $0.10) + (4 x $0.50) = $3.00 + $2.00 = $5.00 a month
Task 3: a batch classification job
100,000 tickets, each with 800 input tokens and 50 output tokens, through each provider's Batch API. That's 80 MTok of input and 5 MTok of output.
- Claude Haiku 5.5 batch ($0.05/$0.25): (80 x $0.05) + (5 x $0.25) = $4.00 + $1.25 = $5.25
- Gemini 3.5 Flash-Lite batch ($0.15/$1.25): (80 x $0.15) + (5 x $1.25) = $12.00 + $6.25 = $18.25
At standard rates the same job is $10.50 on Haiku 5.5 and (80 x $0.30) + (5 x $2.50) = $24.00 + $12.50 = $36.50 on Gemini 3.5 Flash-Lite.
Which API costs less
- Top-tier reasoning: Gemini 4 Carbon is half the price of Opus 5.5 until November 30, then the same.
- Mid-tier volume: Gemini 3.8 Flash costs less per token than Sonnet 5.5, even after its January price rise.
- High-volume, simple tasks: Haiku 5.5 costs less per token than Gemini 3.5 Flash-Lite on prompts up to 100,000 tokens.
- Testing for free: Gemini's free tier, if you accept that your content is used to improve Google's products.
Price per token is only half the bill. A model that finishes a task in fewer tokens or fewer retries can cost less overall, so run your real prompts through both and compare the usage each one reports. If you're choosing coding tools rather than raw APIs, see Claude Code vs Gemini CLI. To start building on Claude, follow our Claude API tutorial.
Learn to build with Claude Code
The Stride AI Coding course teaches Claude Code and Codex for building client tools, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover Claude Code setups, and the Claude Code community is where members compare theirs.
FAQ
How much does the Gemini API cost?
On November 22, 2026, Gemini 4 Carbon is $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 on December 1. Gemini 3.8 Flash is $0.75 and $3.75 until December 31, and Gemini 3.5 Flash-Lite is $0.30 and $2.50.
Is the Gemini API cheaper than the Claude API?
It depends on the tier. Gemini 3.8 Flash costs less per token than Claude Sonnet 5.5, but Claude Haiku 5.5 costs less than Gemini 3.5 Flash-Lite. From December 1, Gemini 4 Carbon and Claude Opus 5.5 cost the same per token.
Does the Gemini API have a free tier?
Yes, on some models, including Gemini 4 Hydrogen, 3.8 Flash and 3.5 Flash-Lite. Google's page says free-tier content is used to improve its products. Gemini 4 Carbon, Argon and Neon aren't on the free tier.
Does the Claude API have a free tier?
No ongoing free tier. New users get free credits to test the API, then pay per token.
Do both APIs offer batch discounts?
Yes. Both list 50% off for batch requests. Claude's batch discount stacks with prompt caching.
Why might my real costs differ from these examples?
Each provider counts tokens differently, models use different numbers of thinking and output tokens for the same task, and caching changes input costs. Measure your own prompts on both before you choose.