The Claude Model Family
Claude models come in tiers. The name tells you size/capability: Haiku (small, fast, cheap), Sonnet (balanced), Opus (large, smart), and the new top tier Fable (most capable). The version number shows the generation (for example 5.5).
2.1 Understanding the tiers
- 1.Haiku: fastest + cheapest; classification, chat, sub-agents
- 2.Sonnet: balance of speed and capability; everyday coding/agents
- 3.Opus: deep reasoning, complex coding; the default choice
- 4.Fable: most capable; hardest, long-horizon tasks
Left to right: both capability and cost go up
2.2 Current models (Oct 2026)
| Model | API model ID | Context | Input $/1M | Output $/1M |
|---|---|---|---|---|
| Claude Fable 5.1 | claude-fable-5-1 | 1M | $10 | $50 |
| Claude Opus 5.5 | claude-opus-5-5 | 1M | $4 | $20 |
| Claude Sonnet 5.5 | claude-sonnet-5-5 | 1M | $2 | $10 |
| Claude Haiku 5.5 | claude-haiku-5-5 | 1M | $0.10 | $0.50 |
| Claude Opus 5 / 4.8 / 4.7 / 4.6 | claude-opus-5, claude-opus-4-8 ... | 1M | $5 | $25 |
| Claude Sonnet 5 | claude-sonnet-5 | 1M | $2 | $10 |
| Claude Haiku 4.5 | claude-haiku-4-5 | 200K | $1 | $5 |
2.3 Which model when?
| Use case | Model | Why |
|---|---|---|
| Complex coding, architecture, agents | Opus 5.5 | Best quality/price; the default |
| Hardest research / multi-hour agents | Fable 5.1 | Maximum intelligence, expensive |
| Daily coding, enterprise chatbot | Sonnet 5.5 | Fast and capable |
| Bulk classification, extraction, routing | Haiku 5.5 | Very cheap and fast |
| Sub-agent / worker that only reads | Haiku or Sonnet | Save cost, keep the lead agent on Opus |
2.4 Model history (short)
| Generation | What was new |
|---|---|
| Claude 1 / 2 (2023) | First public models; Claude 2 went from 100K to 200K context |
| Claude 3 (Mar 2024) | Introduced the Haiku / Sonnet / Opus tiers; vision (images) |
| Claude 3.5 Sonnet (2024) | Came with Artifacts; big jump in coding; computer use (beta) |
| Claude 3.7 Sonnet (Feb 2025) | First hybrid reasoning model (extended thinking); Claude Code preview |
| Claude 4 family (2025) | Opus 4 / Sonnet 4, then 4.1, 4.5 — agentic coding, long tasks |
| Claude 4.6 – 4.8 (2026) | 1M context, adaptive thinking, effort levels, 128K output |
| Claude 5 family (2026) | Opus 5 / 5.5, Sonnet 5 / 5.5, Haiku 5.5, and the new Fable tier |
2.5 How to check a model's capabilities
import anthropic
client = anthropic.Anthropic()
m = client.models.retrieve("claude-opus-5-5")
print(m.display_name, m.max_input_tokens, m.max_tokens) # context window, output cap
for model in client.models.list(): # all available models
print(model.id)