Two vendors changed their AI model prices in the past week, so comparing the cost of running Claude and Gemini through an API means reading two different price tables. Anthropic cut the price of Claude Fable 5.2 by 20%, and Google launched Gemini 4 with a temporary low rate for Carbon and Argon that rises on December 1. This tracker lists the current list prices from each vendor’s own pricing page, what a typical request costs at those rates, and the discounts that change the bill. Prices are per million tokens (MTok), in US dollars, and were read from the vendors’ pricing pages on November 14, 2026.
What does Claude cost through the API?
Anthropic’s pricing page lists these base rates for current models:
| Model | Input | Output | Cache hit |
|---|---|---|---|
| Claude Fable 5.2 | $8 | $40 | $0.20 |
| Claude Mythos 5.2 (limited availability) | $8 | $40 | $0.20 |
| Claude Opus 5.5 | $4 | $20 | $0.20 |
| Claude Sonnet 5.5 | $2 | $10 | $0.10 |
| Claude Haiku 5.5 (prompts up to 100,000 tokens) | $0.10 | $0.50 | $0.01 |
| Claude Haiku 5.5 (prompts over 100,000 tokens) | $0.50 | $2.50 | $0.05 |
Fable 5.1 and the earlier Fable 5 are listed at $10 input and $50 output, which is where the 20% cut comes from. Anthropic also notes that Sonnet 5’s $2 and $10 launch price, first announced as introductory through August 31, is now the standard price; the planned rise to $3 and $15 will not happen.
The page lists several modifiers: the Batch API gives a 50% discount on input and output tokens; fast mode, available for Fable 5.2 and some Opus models, delivers up to 2.5 times higher output speed at premium pricing and cannot be combined with Batch; and specifying US-only inference through the inference_geo parameter applies a 1.1 times multiplier on Claude 4.6 and later models.
What does Gemini 4 cost?
Google’s Gemini API pricing page shows the paid tier, with a rate change on December 1 for two models:
| Model | Input | Output (includes thinking) |
|---|---|---|
| Gemini 4 Carbon, through Nov 30 | $2.00 | $10.00 |
| Gemini 4 Carbon, from Dec 1 | $4.00 | $20.00 |
| Gemini 4 Argon, through Nov 30 | $2.00 | $10.00 |
| Gemini 4 Argon, from Dec 1 | $4.00 | $20.00 |
| Gemini 4 Hydrogen (no scheduled change) | $1.50 | $7.50 |
Per the same page, Batch and Flex processing cost 50% of Standard for all three models, so Hydrogen drops to $0.75 input and $3.75 output. Priority processing costs more: Hydrogen is $2.70 and $13.50, and Carbon is $3.60 and $18.00 at the current rate. Context caching is implicit only, and Hydrogen has a single 1M-token context window with no prompt-length tiers listed.
What does one million tokens in and one million out cost?
To compare the tables, here is the list price of a workload with one million input tokens and one million output tokens, before caching or batch discounts. This is simple arithmetic on the prices above, not a benchmark.
- Claude Fable 5.2: $48
- Claude Opus 5.5: $24
- Claude Sonnet 5.5: $12
- Claude Haiku 5.5 (short prompts): $0.60
- Gemini 4 Carbon: $12 through November 30, then $24
- Gemini 4 Hydrogen: $9 ($4.50 on Batch or Flex)
These are not like-for-like. The models differ in capability, and the bill depends on how many tokens a task uses, including thinking tokens. Google’s own claim that Hydrogen uses less quota than Carbon on a coding benchmark is a vendor figure and does not establish a dollar saving.
What should buyers check before choosing?
- The Carbon and Argon rate change on December 1. Budgets set at $2 and $10 will double for those two models.
- Whether the workload can wait. Both vendors halve the price for batch processing.
- Caching. Anthropic bills cache hits at a fraction of base input; Google caches implicitly.
- Availability. Mythos 5.2 is listed as limited availability, and Gemini 4 Carbon and Argon are not on Google’s free tier.
- Enterprise seats are separate. Anthropic’s plan page lists Team standard seats at $20 per month billed annually ($25 monthly), premium seats at $100 annually ($125 monthly), and Enterprise at $20 per seat per month billed annually with a 20-seat minimum and usage billed at API rates.
OpenAI’s current API rates are not included because its pricing page could not be opened for this report.
Sources
- Anthropic: Claude API pricing
- Anthropic: Claude plans and pricing
- Google: Gemini API pricing
- Sam on AI: AI Brief, November 13, 2026 (Hydrogen pricing and Google’s quota claim)
Image: BalticServers.com / CC BY-SA 3.0 via Wikimedia Commons.
Like our coverage? Add stepphase.com as a preferred source on Google to see more of our stories in Top Stories.
