As of November 9, 2026, Anthropic’s model documentation marks Claude Haiku 5.5 as “latest” in its listed model comparison. Its page lists the API ID claude-haiku-5-5, marks the model active, and includes the Claude API among its platforms. The page gives an October 7, 2026 release date. This is a scoped answer about the documented lineup—not a claim that no newer model exists in a restricted program. Anthropic documents Claude Mythos 5.1 separately as available only to organizations verified through its verification programs.
Latest does not mean best for every workload. Anthropic’s model comparison recommends starting with Claude Opus 5.5 for most workloads; Haiku 5.5 is positioned for high-volume, latency-sensitive tasks. Choose based on your task requirements and evaluation results, not release order alone.
Anthropic API models at a glance
Anthropic’s current model comparison lists four models: Haiku 5.5, Sonnet 5.5, Opus 5.5, and Fable 5.1. Prices below are USD per million tokens (MTok), input/output. Haiku’s starting rates apply to prompts up to 100,000 tokens; longer prompts have higher rates. These are API token prices, not Claude subscription prices.
| Model | Documented positioning | Comparative latency | Input / output price per MTok |
|---|---|---|---|
| Claude Haiku 5.5 | High-volume, latency-sensitive work such as classification, extraction, and routing | Fastest | From $0.10 / $0.50 |
| Claude Sonnet 5.5 | A faster option than Opus in Anthropic’s comparison | Fast | $2 / $10 |
| Claude Opus 5.5 | Anthropic’s suggested starting point for most workloads | Moderate | $4 / $20 |
| Claude Fable 5.1 | A higher-priced option; validate its value on your workload before choosing it | Slower | $10 / $50 |
For Haiku 5.5, Anthropic lists input/output rates of $0.10/$0.50 per MTok for prompts up to 100,000 tokens and $0.50/$2.50 per MTok for prompts over 100,000 tokens. Other usage details, including caching and batch processing, may affect the bill; check the current pricing documentation when estimating costs.
Which Claude API model should you choose?
- Choose Haiku 5.5 for high-volume, latency-sensitive work such as classification, extraction, and routing. Its documented Claude API ID is
claude-haiku-5-5. Before routing production traffic, test quality on representative inputs and account for its higher rates on prompts over 100,000 tokens. - Consider Sonnet 5.5 when its documented “Fast” latency category and $2/$10 per MTok rates fit your application. Compare it with Haiku or Opus on your own tasks; the model comparison alone cannot determine which will perform best for your workload.
- Start with Opus 5.5 if you do not yet know which model fits. Anthropic’s comparison recommends it for most workloads. Its listed rates are higher than Sonnet’s and Haiku’s, so measure task quality and actual token use before scaling.
- Evaluate Fable 5.1 only when your own tests justify its higher listed price and “Slower” latency category. The comparison’s price and latency labels are not, by themselves, evidence that it will deliver better results for your particular task.
A practical selection method is to run representative inputs through the likely candidates, then compare task success, latency, and token use. These are suggested evaluation steps, not reported test results.
What about Claude Mythos 5.1?
Anthropic’s current documentation says Claude Mythos 5.1 is available only to organizations verified through its verification programs, such as the Cyber Verification Program. It lists the API ID claude-mythos-5-1, but that does not make it an ordinary generally available model: access requires verification. Anthropic lists the same specifications and pricing as Fable 5.1.
Separately, TechCrunch reported on April 7, 2026 that an earlier Mythos preview was not then going to be generally available. That is historical reporting about the preview at that time; Anthropic’s current model documentation is the relevant source here for Mythos 5.1’s present documented access condition.
Bottom line
Claude Haiku 5.5 is the model Anthropic’s current comparison marks “latest,” and its model page lists it as active on the Claude API with ID claude-haiku-5-5. For model selection, Haiku is positioned for high-volume, latency-sensitive tasks, while Anthropic recommends Opus as a starting point for most workloads. Mythos 5.1 is a separate, verification-required option—not ordinary open API access. Check Anthropic’s current model and pricing documentation before deployment.
Sources
Claude Haiku 5.5 — Claude Platform Docs
Models overview — Claude Platform Docs