Pricing Comparison
| Model | Provider | Input $/M | Cached $/M | Output $/M | Batch in / out | Context |
|---|---|---|---|---|---|---|
| GPT-6 Luna | OpenAI | $0.10 | $0.01 | $0.50 | $0.05 / $0.25 | 1.05M |
| Claude Haiku 5.5 | Anthropic | $0.10 | $0.01 | $0.50 | $0.05 / $0.25 | 1M |
| GPT-4o miniDeprecated | OpenAI | $0.15 | $0.075 | $0.60 | $0.075 / $0.30 | 128K |
| Mistral Small 4 | Mistral | $0.15 | $0.015 | $0.60 | β | 256K |
| Gemini 3.5 Flash-Lite | $0.30 | $0.03 | $2.50 | $0.15 / $1.25 | 1M | |
| Gemini 2.5 Flash | $0.30 | $0.03 | $2.50 | $0.15 / $1.25 | 1M | |
| Mistral Large 3 | Mistral | $0.50 | $0.05 | $1.50 | β | 256K |
| Mistral Large 4 | Mistral | $0.68 | $0.07 | $2.09 | β | 1M |
| Gemini 3.8 Flash | $0.75 | $0.075 | $3.75 | $0.375 / $1.88 | 1M | |
| Claude Haiku 4.5 | Anthropic | $1.00 | $0.10 | $5.00 | $0.50 / $2.50 | β |
| o3-miniDeprecated | OpenAI | $1.10 | $0.55 | $4.40 | $0.55 / $2.20 | 200K |
| GPT-5Deprecated | OpenAI | $1.25 | $0.125 | $10.00 | $0.625 / $5.00 | 400K |
| Gemini 2.5 Pro | $1.25 | $0.125 | $10.00 | $0.625 / $5.00 | 1M | |
| Grok 4.3 | xAI | $1.25 | $0.20 | $2.50 | β | 1M |
| Mistral Medium 3.5 | Mistral | $1.50 | $0.15 | $7.50 | β | 256K |
| GPT-6.1 Sol | OpenAI | $2.00 | $0.10 | $10.00 | $1.00 / $5.00 | 1.05M |
| o3Deprecated | OpenAI | $2.00 | $0.50 | $8.00 | $1.00 / $4.00 | 200K |
| Claude Sonnet 5.5 | Anthropic | $2.00 | $0.10 | $10.00 | $1.00 / $5.00 | 1M |
| Claude Sonnet 5 | Anthropic | $2.00 | $0.20 | $10.00 | $1.00 / $5.00 | 1M |
| Gemini 4 Argon | $2.00 | $0.10 | $10.00 | $1.00 / $5.00 | 1M | |
| Grok 4.7 | xAI | $2.00 | $0.50 | $6.00 | β | 500K |
| GPT-4oDeprecated | OpenAI | $2.50 | $1.25 | $10.00 | $1.25 / $5.00 | 128K |
| Claude Opus 5.5 | Anthropic | $4.00 | $0.20 | $20.00 | $2.00 / $10.00 | 1M |
| Claude Opus 5 | Anthropic | $5.00 | $0.50 | $25.00 | $2.50 / $12.50 | 1M |
| GPT-6 Astra | OpenAI | $10.00 | $1.00 | $50.00 | $5.00 / $25.00 | 1.05M |
| Claude Fable 5.1 | Anthropic | $10.00 | $0.25 | $50.00 | $5.00 / $25.00 | 1M |
Standard-tier list prices in USD per million tokens via each provider's own API, as of October 16, 2026. βCachedβ is the cache-hit (cached input) rate; cache writes and storage fees are extra where they apply. βββ means the provider's pricing page does not list that rate. Long-context surcharges are in the notes. Cloud marketplaces (Bedrock, Vertex, Azure) can price differently.
Input Cost per Million Tokens
Output Cost per Million Tokens
Cost Calculator
Estimate your monthly AI spend across models
| Model | Monthly Cost |
|---|---|
| GPT-6 LunaOpenAI | $0.30 |
| Claude Haiku 5.5Anthropic | $0.30 |
| GPT-4o miniOpenAIDeprecated | $0.38 |
| Mistral Small 4Mistral | $0.38 |
| Mistral Large 3Mistral | $1.00 |
| Mistral Large 4Mistral | $1.39 |
| Gemini 3.5 Flash-LiteGoogle | $1.40 |
| Gemini 2.5 FlashGoogle | $1.40 |
| Grok 4.3xAI | $1.88 |
| Gemini 3.8 FlashGoogle | $2.25 |
| o3-miniOpenAIDeprecated | $2.75 |
| Claude Haiku 4.5Anthropic | $3.00 |
| Grok 4.7xAI | $4.00 |
| Mistral Medium 3.5Mistral | $4.50 |
| o3OpenAIDeprecated | $5.00 |
| GPT-5OpenAIDeprecated | $5.63 |
| Gemini 2.5 ProGoogle | $5.63 |
| GPT-6.1 SolOpenAI | $6.00 |
| Claude Sonnet 5.5Anthropic | $6.00 |
| Claude Sonnet 5Anthropic | $6.00 |
| Gemini 4 ArgonGoogle | $6.00 |
| GPT-4oOpenAIDeprecated | $6.25 |
| Claude Opus 5.5Anthropic | $12.00 |
| Claude Opus 5Anthropic | $15.00 |
| GPT-6 AstraOpenAI | $30.00 |
| Claude Fable 5.1Anthropic | $30.00 |
*Assumes ~4K tokens per request average. Uses standard short-context rates; caching, batch, and long-context pricing are not applied.
Provider Deep Dive
OpenAI
GPT-6 Astra ($10/$50) is the flagship; GPT-6.1 Sol ($2/$10, $0.10 cached) is the mid tier; GPT-6 Luna ($0.10/$0.50) is the budget tier. All three list a 1.05M context window and charge higher rates above 272K input tokens. Batch is half price. GPT-4o and GPT-4o mini were deprecated Oct 14, 2026 and leave the API Jan 14, 2027.
Anthropic
Four current tiers: Fable 5.1 ($10/$50), Opus 5.5 ($4/$20), Sonnet 5.5 ($2/$10), and Haiku 5.5 ($0.10/$0.50, rising to $0.50/$2.50 for prompts over 100K tokens). All have 1M context windows. Batch is 50% off, and cache hits on Opus 5.5 and Sonnet 5.5 cost 5% of the input price.
Gemini 4 Argon is $2/$10 through Nov 30, 2026, then $4/$20 from Dec 1. Gemini 3.8 Flash is $0.75/$3.75 through Dec 31, then $1.50/$7.50. Gemini 3.5 Flash-Lite is $0.30/$2.50. Gemini 2.5 Pro and 2.5 Flash remain available with no shutdown date announced. All listed models have ~1M context.
xAI
Grok 4.7 is the flagship at $2/$6 per million tokens with a 500K context window, doubling to $4/$12 once a prompt reaches 200K tokens. Grok 4.3 is $1.25/$2.50 with a 1M context window.
Mistral
Mistral Large 4 (1M context) is on sale at $0.68/$2.09, down from $1.36/$4.18; Mistral does not state an end date. Large 3 is $0.50/$1.50, Medium 3.5 is $1.50/$7.50, and Small 4 is $0.15/$0.60, each with 256K context.
Pricing Trends
What the current price sheets show
Three-Tier Lineups Are Standard
OpenAI (GPT-6 Astra / 6.1 Sol / 6 Luna) and Anthropic (Fable 5.1 / Sonnet 5.5 / Haiku 5.5) now sell the same ladder: a $10/$50 top tier, a $2/$10 workhorse, and a $0.10/$0.50 budget model. Anthropic adds Opus 5.5 at $4/$20 in between.
Caching and Batch Are Where the Savings Are
Batch APIs are 50% off at OpenAI, Anthropic, and Google. On the newest OpenAI, Anthropic, and Google models, cache hits cost 2.5β10% of the base input price, so a repeated system prompt or document is far cheaper than the headline rate.
Long Prompts Cost More
Several providers now charge more per token past a threshold: OpenAI above 272K input tokens, xAI at 200K, Gemini 2.5 Pro above 200K, and Claude Haiku 5.5 above 100K. Most Claude models and the newest Gemini models charge one rate across the full 1M window.
Watch the Promo Dates
Gemini 4 Argon's price doubles on Dec 1, 2026 and Gemini 3.8 Flash's doubles on Jan 1, 2027. Mistral Large 4 is on a sale with no stated end date. Budget against the regular price.
Common Questions
Which AI model is cheapest per token?
Among current, non-deprecated models on this page, GPT-6 Luna and Claude Haiku 5.5 tie for the lowest input price at $0.10 per million tokens, both with $0.50 output. Haiku 5.5 rises to $0.50/$2.50 for prompts over 100K tokens. Mistral Small 4 ($0.15/$0.60) is next.
How do I choose between models?
Match cost to task complexity. Use the budget tier ($0.10β$0.75 input) for classification, extraction, and simple generation. Use the $2/$10 tier (GPT-6.1 Sol, Claude Sonnet 5.5, Gemini 4 Argon) for coding and analysis. Reserve the top tier ($4β$10 input) for work where quality directly changes the outcome. Test on your own prompts, since token counts differ by tokenizer.
What do cached input and batch pricing mean?
Cached input is the rate for prompt tokens the provider has already cached, such as a repeated system prompt; on current OpenAI, Anthropic, and Google models it is 2.5β10% of the normal input price (older models such as GPT-4o charge 50%). Batch pricing applies when you submit requests asynchronously through a batch API and accept delayed results; OpenAI, Anthropic, and Google charge half price.
Which models are being deprecated?
OpenAI deprecated GPT-4o and GPT-4o mini on Oct 14, 2026, with API removal on Jan 14, 2027. The gpt-5-2025-08-07 and o3-2025-04-16 snapshots shut down Dec 11, 2026, and o3-mini shuts down Oct 23, 2026. Google lists no shutdown date for Gemini 2.5 Pro or 2.5 Flash.