ModelsPricing history
Model pricing
Per-token list prices over time — input and output USD per million tokens, the same convention as Models. Pick models to compare. Steps hold until the next sourced price change; we do not invent a series when only the current list price is known.
The seed mixes well-documented earlier generations (GPT-4 → GPT-4o, Claude 3 → 3.5, early Gemini Flash/Pro bands) with the current flagship cut from the Models page. Cache discounts, batch rates, and over-context multipliers are noted in the table, not plotted.
For a current tier comparison — list $/1M next to context and value attributes — see Price vs value.
Not investment advice. Figures lag the labs. Prefer lab pricing pages and announcements; Artificial Analysis fills gaps where we measure first-party API list prices for open-weight hosts.
Compare models
Toggle models to plot input and output list prices over time.
USD per 1M tokens · step holds until next priced date
USD per 1M tokens · step holds until next priced date
Price points
Rows for the models currently selected. Newest first.
| As of | Model | Input $/1M | Output $/1M | Source | Notes |
|---|---|---|---|---|---|
| Aug 6, 2024 | GPT-4o | $2.50 | $10 | Source | Published list-price cut (~50% input). Some secondary trackers date the cut later in 2024; use OpenAI pricing page / changelog as authority. |
| Jun 20, 2024 | Claude 3.5 Sonnet | $3 | $15 | Source | Launch list price matched Claude 3 Sonnet. |
| May 14, 2024 | Gemini 1.5 Pro | $3.50 | $10.50 | Source | Commonly cited mid-2024 Google AI Studio / Vertex list for ≤128k prompts; Google later cut Flash/Pro bands — treat as a period snapshot, not a continuous series. |
| May 13, 2024 | GPT-4o | $5 | $15 | Source | Launch list price; half of GPT-4 Turbo input. |
| Mar 4, 2024 | Claude 3 Opus | $15 | $75 | Source | Claude 3 family launch list price. |
| Mar 14, 2023 | GPT-4 (8K) | $30 | $60 | Source | Launch list price for gpt-4 (8K). Widely archived as $0.03/$0.06 per 1K tokens (= $30/$60 per 1M). |