Pricing milestones
| Date | Pricing event | What to cite |
|---|---|---|
| 2023-07-11 | Claude 2 API was offered at the same price as Claude 1.3. | Claude 2 launch |
| 2023-11-21 | Claude 2.1 announced updated pricing to improve cost efficiency across models. | Claude 2.1 launch |
| 2024-06-21 | Claude 3.5 Sonnet launched at $3 input / $15 output per MTok. | Claude 3.5 Sonnet |
| 2025-09-29 | Claude Sonnet 4.5 kept Sonnet 4 pricing at $3/$15 per MTok. | Sonnet 4.5 |
| 2025-11-24 | Claude Opus 4.5 launched at $5/$25 per MTok. | Opus 4.5 |
| 2026-06-30 | Claude Sonnet 5 launched with introductory $2/$10 per MTok through August 31, 2026, then $3/$15. | Pricing docs |
Current rate snapshot
As of the July 2026 pricing docs, current headline rates include Fable 5 and Mythos 5 at $10 input / $50 output per MTok, Opus 4.8/4.7/4.6/4.5 at $5/$25, Sonnet 5 at $2/$10 through August 31, 2026 then $3/$15, Sonnet 4.6 and 4.5 at $3/$15, and Haiku 4.5 at $1/$5.
The same page records prompt-caching columns, batch processing discounts, fast mode, data-residency multipliers, and partner-platform caveats. Those modifiers make the official pricing page the safer citation than a copied table on an old blog post.
The real pricing history is context plus tokens
Large context windows changed what one request could contain. Claude 2 publicized 100K tokens, Claude 2.1 moved to 200K, and modern Fable/Opus/Sonnet entries list 1M-token context windows. That makes price history inseparable from context-window history: the same per-token rate can support very different product designs if the model can keep more material in one request.
Tokenizers also matter. The pricing docs say Claude Opus 4.7 and later Opus models, Fable 5, Mythos 5, Mythos Preview, and Sonnet 5 use a newer tokenizer that produces approximately 30% more tokens for the same text than Sonnet 4.6 and earlier. That is a pricing-history fact, not only a migration footnote.
Avoid stale pricing quotes
Pricing pages change. Launch posts are useful for historical rates, but current budget decisions should cite the live pricing docs. When a page here mentions a rate, it names the source date and links directly to the official Anthropic source.
Questions this page answers
What was Claude 3.5 Sonnet priced at launch?
Anthropic stated Claude 3.5 Sonnet cost $3 per million input tokens and $15 per million output tokens at launch.
What is Claude Sonnet 5 introductory pricing?
The official pricing docs list introductory $2 input / $10 output per MTok through August 31, 2026, then $3/$15.
Why can the same text cost differently on newer Claude models?
The pricing docs say Sonnet 5 and newer Opus/Fable/Mythos models use a newer tokenizer that can produce approximately 30% more tokens for the same text.