Anthropic · Frontier
Checked as of 2026-10-23
Claude
Anthropic's models: one family from small and fast to frontier reasoning, all with a 1M-token context.
At a glance
- Licence
- Proprietary Anthropic Commercial Terms
Closed weights. API use under Anthropic's Commercial Terms; on Bedrock or Google Cloud, the cloud's terms.
- Access
- APIBedrockVertex AIAzure
Variants 5
| Model | Context | $ in / 1M | $ out / 1M | Inputs | Status |
|---|---|---|---|---|---|
Claude Fable 5.1
claude-fable-5-1
For the hardest reasoning and long-horizon agent work. Thinking always on. |
1M | $10 | $50 | text, image | current |
Claude Opus 5.5
claude-opus-5-5
Anthropic's recommended starting point for most workloads. |
1M | $4 | $20 | text, image | current |
Claude Sonnet 5.5
claude-sonnet-5-5
|
1M | $2 | $10 | text, image | current |
Claude Haiku 5.5
claude-haiku-5-5
Prompts over 100K tokens are billed at $0.50 / $2.50. |
1M | $0.1 | $0.5 | text, image | current |
Claude Opus 5
claude-opus-5
Still available; Anthropic recommends moving to Opus 5.5, which is cheaper. |
1M | $5 | $25 | text, image | legacy |
"—" means the figure could not be verified from a primary source on 2026-10-23 and has been left blank rather than guessed. Prices are USD per million tokens at the standard tier.
Strengths
- Every current model has a 1M-token context and 128K output. Fable, Opus and Sonnet bill the full 1M at the standard rate.
- The widest cloud reach of the three: Anthropic's API, Amazon Bedrock, Google Cloud and Microsoft Foundry, under the same model IDs.
- Deep discounts for prompt caching (cache reads 2.5–10% of the input price) and 50% off through the Batch API.
- Published retirement dates: current models no sooner than September–October 2027.
Watch out for
- Inputs are text and images only. No native audio or video.
- Opus 5.5 and Fable 5.1 always think, and forced tool use returns an error. Code written for older models may need changes.
- Haiku 5.5 is five times dearer for prompts over 100K tokens.
- The tokenizer since Opus 4.7 produces about 30% more tokens for the same text, so compare costs in dollars, not tokens.