Anthropic AI Models: Every Claude Model and Which to Use
Anthropic AI models compared: Fable 5.2, Opus 5.5, Sonnet 5.5 and Haiku 5.5 with IDs, prices and context, plus Claude Code's default model and how to switch.
Josh Pocock7 min read
AI Coding
Claude models
Stride AI Academy
Anthropic's current AI models are Claude Fable 5.2, Claude Opus 5.5, Claude Sonnet 5.5 and Claude Haiku 5.5. All four have a 1M-token context window, 128K max output and a June 2026 knowledge cutoff. Anthropic's advice is to start with Opus 5.5 for most work, move up to Fable 5.2 for the hardest reasoning and long agent runs, and use Sonnet or Haiku when speed and cost matter more. Claude Code uses Opus 5.5 by default on every paid plan.
This guide covers the current lineup with model IDs and prices, what each model is best at, the older models still on offer, and which model Claude Code uses and how to switch. Everything here was checked against Anthropic's models overview, pricing docs, Claude's pricing page and the Claude Code model configuration docs on November 22, 2026. Anthropic ships new models often, so check those pages if the version numbers have moved on.
The current Claude AI models
From the models overview, November 22, 2026:
| Claude Fable 5.2 | Claude Opus 5.5 | Claude Sonnet 5.5 | Claude Haiku 5.5 | |
|---|---|---|---|---|
| Best for | Demanding reasoning and long-horizon agentic work | Long-running agentic coding and knowledge work | The best mix of speed and intelligence | High-volume, latency-sensitive tasks |
| API model ID | claude-fable-5-2 |
claude-opus-5-5 |
claude-sonnet-5-5 |
claude-haiku-5-5 |
| Speed | Moderate | Moderate | Fast | Fastest |
| Input price (per MTok) | $8 | $4 | $2 | From $0.10 |
| Output price (per MTok) | $40 | $20 | $10 | From $0.50 |
| Thinking | Adaptive, always on | Adaptive, always on | Adaptive | Adaptive |
| Default effort (API) | high |
medium |
high |
medium |
| Context window | 1M tokens | 1M tokens | 1M tokens | 1M tokens |
| Max output | 128K tokens | 128K tokens | 128K tokens | 128K tokens |
| Knowledge cutoff | June 2026 | June 2026 | June 2026 | June 2026 |
MTok means a million tokens. Haiku 5.5 is $0.10 input and $0.50 output for prompts up to 100K tokens, and $0.50 and $2.50 above that. All four take text and image input and support vision, tool use and multiple languages.
What each model is good at
Claude Opus 5.5. The default pick. Anthropic's docs say to start here if you're unsure. It's built for long agentic coding sessions and knowledge work, and it's what Claude Code runs unless you change it.
Claude Fable 5.2. The most capable tier and twice Opus's price. The docs suggest it for demanding reasoning and long-horizon agent work, or when Opus 5.5 at higher effort still falls short in your tests. On the Claude apps it's limited by plan: Pro pays for it with usage credits, and Max and Ultra can spend up to half their weekly limit on it.
Claude Sonnet 5.5. Half Opus's price and faster. A strong fit for everyday coding, writing and agents where you want quick turns. If you're weighing the two, our Sonnet vs Opus comparison goes deeper.
Claude Haiku 5.5. The fastest and cheapest. Use it for classification, extraction, routing and other high-volume jobs where each call is small.
Older Claude models still available
The overview lists older models you can still call, though Anthropic says to default to the latest where possible:
| Model | Context | Input / output (per MTok) |
|---|---|---|
| Claude Fable 5.1, Fable 5 | 1M | $10 / $50 |
| Claude Opus 5, Opus 4.8, 4.7, 4.6 | 1M | $5 / $25 |
| Claude Opus 4.5 | 200K | $5 / $25 |
| Claude Sonnet 5 | 1M | $2 / $10 |
| Claude Sonnet 4.6 | 1M | $3 / $15 |
| Claude Sonnet 4.5 | 200K | $3 / $15 |
| Claude Haiku 4.5 | 200K | $1 / $5 |
Claude Mythos 5.2, 5.1 and 5 are listed with limited availability: only organizations verified through Anthropic's verification programs can use them. Opus 4.1, Opus 4 and Sonnet 4 are retired on Anthropic's own platform.
One thing to know when comparing old and new prices: the docs say the tokenizer introduced with Opus 4.7 fits about 555,000 words into 1M tokens, against about 750,000 on earlier models. The same text costs more tokens on newer models.
Claude model list on the API
The model ID is what you send in the model field. Every Claude model ID is a pinned snapshot, including the dateless IDs used from the 4.6 generation on, so claude-opus-5-5 won't silently change under you.
import anthropic
client = anthropic.Anthropic()
message = client.messages.create(
model="claude-sonnet-5-5",
max_tokens=1024,
messages=[{"role": "user", "content": "Summarize this ticket in two sentences."}],
)
print(message.content)To pull the live list with each model's token limits and capabilities, call the Models API. Each entry includes max_input_tokens, max_tokens and a capabilities object.
The same models run on Amazon Bedrock, Google Cloud, Microsoft Foundry and Claude Platform on AWS, with platform-specific IDs listed in the overview. For costs beyond the base price, including caching and batch discounts, see Claude API pricing. Our Claude API tutorial covers the first call end to end.
Which models Claude Code uses
Per the model configuration docs, Claude Code's default is Opus 5.5 on Pro, Max, Ultra, Team and Enterprise plans and with an Anthropic API key. Microsoft Foundry is the exception, where the default is Sonnet 4.5. No Fable model is the default on any plan. An admin-set organization default overrides all of these. The Free plan doesn't include Claude Code.
Claude Code also sets its own default effort, which differs from the API's: medium for Fable 5.2, Opus 5.5, Sonnet 5.5 and Haiku 5.5.
Claude Code model aliases
| Alias | What you get |
|---|---|
opus or best |
Latest Opus (Opus 5.5 on the Anthropic API) |
sonnet |
Latest Sonnet (Sonnet 5.5) |
haiku |
Latest Haiku (Haiku 5.5) |
fable |
Fable 5.2 |
opusplan |
Opus in plan mode, then Sonnet for execution |
default |
Clears your override and uses your account's default |
Aliases can point to older versions on some cloud providers. To pin a version, use the full ID, such as claude-opus-5-5.
How to switch models in Claude Code
During a session, open the picker or name the model:
/model sonnetIn the picker, Enter switches and saves the choice as your default, and s switches for this session only.
At startup:
claude --model opusIn your settings file, which our settings.json guide covers in full:
{
"model": "opus"
}Or with an environment variable for one session:
export ANTHROPIC_MODEL=sonnetWhen more than one is set, the docs give this order, highest first: /model in the session, then --model, then ANTHROPIC_MODEL, then the model setting, then ANTHROPIC_DEFAULT_MODEL.
Fable has to be picked explicitly with /model fable or claude --model fable, and it needs Claude Code v2.1.321 or later. On Pro, Fable usage bills to usage credits rather than your plan limit. Our Claude Code usage guide and Claude usage limits explain how model choice affects how fast you hit a limit.
Change effort instead of model
Often the cheaper fix is effort, not a new model. Set it with:
/effort highCurrent models accept low, medium, high, xhigh and max. Lower effort means less thinking and fewer tokens.
Learn to build with Claude Code
The Stride AI Coding course teaches Claude Code and Codex for building client tools, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover Claude Code setups, and the Claude Code community is where members compare theirs.
FAQ
What are Anthropic's current AI models?
On November 22, 2026: Claude Fable 5.2, Claude Opus 5.5, Claude Sonnet 5.5 and Claude Haiku 5.5. All have a 1M-token context window and 128K max output, and Anthropic suggests starting with Opus 5.5.
Which Claude model is best?
Fable 5.2 is the most capable and the most expensive at $8 and $40 per million tokens. For most work Anthropic recommends Opus 5.5, with Sonnet 5.5 for speed and Haiku 5.5 for high-volume tasks.
What model does Claude Code use by default?
Opus 5.5 on Pro, Max, Ultra, Team and Enterprise plans and with an Anthropic API key. Microsoft Foundry defaults to Sonnet 4.5. No plan defaults to Fable.
How do I change the model in Claude Code?
Run /model and pick one, or type /model sonnet. You can also start with claude --model opus, set ANTHROPIC_MODEL, or add a model key to your settings file.
What's the context window of Claude models?
All four current models have a 1M-token context window and 128K tokens of max output. Some older models, such as Opus 4.5, Sonnet 4.5 and Haiku 4.5, have 200K.
What is Claude Mythos?
Mythos 5.2, 5.1 and 5 are Claude models with limited availability. Only organizations verified through Anthropic's verification programs, such as its Cyber Verification Program, can use them.