WORKLOAD INPUT
Request assumptions.
Direct token charges only. Tool calls, storage, regional pricing and non-Standard tiers are excluded.
MODEL ECONOMICS / VERIFIED 2026-09-27
Model pricing.
Normalized.
One source-backed view of Standard API token pricing and core limits across selected frontier models. Compare rates, inspect pricing rules and estimate a request without mixing incompatible service tiers.
PRICING DATABASE
Compare the published baseline.
Rates below reflect the catalog's verified Standard period. Model-specific long-context or scheduled-rate changes are called out in the model row.
| Model | Provider | Context | Max output | Input / MTok | Cached / MTok | Output / MTok | Source |
|---|---|---|---|---|---|---|---|
| GPT-6 Astra gpt-6-astra Long context > 272,000: 2× input/cache · 1.5× output | OpenAI | 1,050,000tokens | 128,000tokens | $10.00 | $1.00 | $50.00 | Official ↗ |
| GPT-6 Sol gpt-6-sol Long context > 272,000: 2× input/cache · 1.5× output | OpenAI | 1,050,000tokens | 128,000tokens | $2.00 | $0.20 | $10.00 | Official ↗ |
| GPT-6 Luna gpt-6-luna Long context > 272,000: 2× input/cache · 1.5× output | OpenAI | 1,050,000tokens | 128,000tokens | $0.10 | $0.01 | $0.50 | Official ↗ |
| GPT-5.6 Sol gpt-5.6-sol Long context > 272,000: 2× input/cache · 1.5× outputCurrent $4/$20 Standard rates are promotional through at least November 21, 2026. | OpenAI | 1,050,000tokens | 128,000tokens | $4.00 | $0.40 | $20.00 | Official ↗ |
| GPT-5.6 Terra gpt-5.6-terra Long context > 272,000: 2× input/cache · 1.5× output | OpenAI | 1,050,000tokens | 128,000tokens | $2.00 | $0.20 | $12.00 | Official ↗ |
| GPT-5.6 Luna gpt-5.6-luna Long context > 272,000: 2× input/cache · 1.5× output | OpenAI | 1,050,000tokens | 128,000tokens | $0.20 | $0.02 | $1.20 | Official ↗ |
| Claude Fable 5.1 claude-fable-5-1 | Anthropic | 1,000,000tokens | 128,000tokens | $10.00 | $0.25 | $50.00 | Official ↗ |
| Claude Opus 5.5 claude-opus-5-5 | Anthropic | 1,000,000tokens | 128,000tokens | $4.00 | $0.20 | $20.00 | Official ↗ |
| Claude Sonnet 5 claude-sonnet-5 | Anthropic | 1,000,000tokens | 128,000tokens | $2.00 | $0.20 | $10.00 | Official ↗ |
| Claude Haiku 4.5 claude-haiku-4-5-20251001 | Anthropic | 200,000tokens | 64,000tokens | $1.00 | $0.10 | $5.00 | Official ↗ |
| Gemini 3.8 Flash gemini-3.8-flash From 2027-01-01: $1.50 input · $7.50 outputIntroductory Standard pricing applies through December 31, 2026; higher Standard pricing starts January 1, 2027. | 1,048,576tokens | 65,536tokens | $0.75 | $0.075 | $3.75 | Official ↗ |
No models match this filter.
Catalog verified 2026-09-27. Promotional or scheduled rates can change; the official vendor documentation remains the final billing authority.
Machine-readable JSON ↗
TOKEN COST CALCULATOR
Estimate one workload.
Enter uncached input, cached input and output tokens. The calculator selects the Standard pricing period by date and applies stored long-context rules automatically.
ESTIMATED DIRECT TOKEN COST
Standard API estimate.
Calculated from the same public catalog used to render the table above.
ESTIMATED TOTAL—
Uncached input—
Cached input—
Output—
Loading pricing dataset…
METHODOLOGY
Comparable first. Caveats visible.
SXF separates directly comparable Standard token rates from service tiers and add-ons that can change the bill but are not equivalent across providers.
One unit.
Headline token rates are normalized to USD per one million tokens. Cached input stays separate from ordinary input.
Request shape matters.
Stored long-context rules and dated pricing schedules are applied from the canonical dataset instead of being retyped into the calculator.
Do not mix tiers.
Batch, Flex, Fast, Priority, tools, grounding, storage, regional uplifts and negotiated pricing remain outside the headline calculator.
PRICING FAQ
How to read the database.
What pricing does the SXF model database use?
The table normalizes vendor-listed Standard API token pricing in USD per one million tokens. It does not mix Batch, Flex, Fast, Priority, regional, enterprise, or negotiated pricing into the headline rates.
Does the calculator include tool or grounding charges?
No. The calculator estimates direct text-token charges only. Search, grounding, code execution, tools, cache storage, regional uplifts, and other add-on charges are excluded unless they are explicitly represented as token rates.
How does long-context pricing work for OpenAI models in this database?
For cataloged OpenAI models with a long-context rule, requests above 272,000 total input tokens use the published higher input, cache, and output rates for the full request. The calculator applies that rule automatically.
Why does Gemini 3.8 Flash show a future price change?
Google published introductory Standard pricing through December 31, 2026 and higher Standard pricing beginning January 1, 2027. The calculator selects the stored pricing period from the billing date.