GPT-6 Luna
Oct 6, 2026OpenAI
- Input / 1M
- USD 0.1
- Output / 1M
- USD 0.5
- Cached input
- USD 0.01
- Context
- 1,050,000
1 mapped APIModel details →
Reference view
LLM API Pricing
A graph-derived comparison of standard token pricing for indexed language models with published pricing data. API availability and providers are resolved from verified relationships.
Prices are shown per one million tokens in the model record's currency. Cache writes, batch discounts, long-context tiers, regional processing, tool calls, and promotional rates may follow separate rules; open the model page and its primary sources before making a purchasing decision.
Priced models
13
Indexed models
34
Without pricing
21
Verification window
Oct 3, 2026 – Oct 6, 2026
Pricing table
Published token prices
GPT-5.6 Luna
Oct 4, 2026OpenAI
- Input / 1M
- USD 0.2
- Output / 1M
- USD 1.2
- Cached input
- USD 0.02
- Context
- 1,050,000
1 mapped APIModel details →
Gemini 3.8 Flash
Oct 3, 2026Google
- Input / 1M
- USD 0.75
- Output / 1M
- USD 3.75
- Cached input
- USD 0.075
- Context
- 1,000,000
1 mapped APIModel details →
Mistral Medium 3.5
Oct 3, 2026Mistral AI
- Input / 1M
- USD 1.5
- Output / 1M
- USD 7.5
- Cached input
- USD 0.15
- Context
- 256,000
1 mapped APIModel details →
Claude Sonnet 5.5
Oct 3, 2026Anthropic
- Input / 1M
- USD 2
- Output / 1M
- USD 10
- Cached input
- USD 0.2
- Context
- 1,000,000
1 mapped APIModel details →
GPT-5.6 Terra
Oct 3, 2026OpenAI
- Input / 1M
- USD 2
- Output / 1M
- USD 12
- Cached input
- USD 0.2
- Context
- 1,050,000
1 mapped APIModel details →
GPT-6 Sol
Oct 6, 2026OpenAI
- Input / 1M
- USD 2
- Output / 1M
- USD 10
- Cached input
- USD 0.2
- Context
- 1,050,000
1 mapped APIModel details →
GPT-6.1 Sol
Oct 3, 2026OpenAI
- Input / 1M
- USD 2
- Output / 1M
- USD 10
- Cached input
- USD 0.1
- Context
- 1,050,000
1 mapped APIModel details →
Grok 4.7
Oct 3, 2026SpaceXAI
- Input / 1M
- USD 2
- Output / 1M
- USD 6
- Cached input
- —
- Context
- 500,000
1 mapped APIModel details →
Claude Opus 5.5
Oct 3, 2026Anthropic
- Input / 1M
- USD 4
- Output / 1M
- USD 20
- Cached input
- USD 0.2
- Context
- 1,000,000
1 mapped APIModel details →
GPT-5.6 Sol
Oct 3, 2026OpenAI
- Input / 1M
- USD 4
- Output / 1M
- USD 20
- Cached input
- USD 0.4
- Context
- 1,050,000
1 mapped APIModel details →
Claude Fable 5.1
Oct 3, 2026Anthropic
- Input / 1M
- USD 10
- Output / 1M
- USD 50
- Cached input
- USD 0.25
- Context
- 1,000,000
1 mapped APIModel details →
GPT-6 Astra
Oct 3, 2026OpenAI
- Input / 1M
- USD 10
- Output / 1M
- USD 50
- Cached input
- USD 1
- Context
- 1,050,000
1 mapped APIModel details →
Turn pricing into a workload estimate
What would your actual API workload cost?
Use the same source-backed token prices with your own input, output, cache, request-count or budget assumptions.
How to read this table
This view intentionally avoids converting everything into a single score or declaring a cheapest model. Pricing is only one dimension; context limits, output limits, modalities, tool support, provider availability, and workload-specific pricing rules can materially change the effective cost.