Skip to main content
Using Claude

Claude Model Selection Guide: Opus 5 vs Sonnet 5 vs Haiku 4.5

A side-by-side comparison of Claude Opus 5, Sonnet 5, and Haiku 4.5 — and a clear guide to picking the right model for your task.

ByUpdated
🌐 This article was machine-translated and may contain inaccuracies. Read the Korean original if in doubt.

Claude offers three models with different strengths — Opus 5, Sonnet 5, and Haiku 4.5. The smartest model is not always the right answer. Pick based on the task's difficulty, speed, and cost. This guide lays out how the three differ and when to use which.

🟢 Model references match the current lineup · model notice · Fable subscription
🟢 Model references match the current lineup · Claude Opus 5.5 / Claude Sonnet 5 / Claude Haiku 4.5 (higher tier: Claude Fable 5.1). This notice changes only when Anthropic ships a new model.

Fable 5 and 5.1 subscription (updated September 7, 2026): Claude Fable 5.1, released September 1, 2026, is the current Fable model and Fable 5 is now legacy. Plan terms are the same for both — Max and Team Premium plans include Fable at up to 50% of the weekly usage limit; Pro and Team Standard use usage credits (

🟢 Model references match the current lineup · Claude Opus 5.5 / Claude Sonnet 5 / Claude Haiku 4.5 (higher tier: Claude Fable 5.1). This notice changes only when Anthropic ships a new model.
0/M input, $50/M output tokens). The one-time
🟢 Model references match the current lineup · Claude Opus 5.5 / Claude Sonnet 5 / Claude Haiku 4.5 (higher tier: Claude Fable 5.1). This notice changes only when Anthropic ships a new model.
00 credit applied only to the Fable 5 transition and is not offered for 5.1. Some coding and debugging requests may be answered by an Opus model due to a security classifier (both models). See the Fable 5.1 guide and the Fable 5 availability guide for details.

Update (September 27, 2026): Claude Opus 5.5 was released on September 22, 2026 and is now the current Opus model (API claude-opus-5-5, $4 per million input tokens and $20 per million output tokens, down from Opus 5's $5/$25). Anthropic says it performs at the level of Fable 5.1 on most work and costs 40% less to run than Opus 5 on typical workloads. Opus 5 remains available but is marked legacy in the official model docs. Opus 5 prices and comparisons below reflect the pre-launch state and will be updated. Details: Claude Opus 5.5 complete guide.

Models at a glanceOpus 5Most capable — hard workContext 1M / Output 128K$5 / $25 (per 1M tokens)Complex reasoning, long codingSonnet 5Balanced — enhanced agenticContext 1M / Output 128K$2 / $10 (standard price)Sep 1 increase cancelledDefault on Free & Pro plansHaiku 4.5Fastest — lightweightContext 200K / Output 64K$1 / $5 (per 1M tokens)Fast, high-volume, low-cost

The differences at a glance

Here are the core specs (Claude API, per official docs). Pricing is input / output per 1M tokens — a token being the unit Claude uses to process text; roughly one short word in English.

  • Opus 5 — most capable. 1M context / 128K output tokens. $5 / $25.
  • Sonnet 5 — balanced, enhanced agentic performance. 1M / 128K tokens. $2 / $10 — now the standard price. It was announced as introductory pricing through August 31, 2026, but Anthropic cancelled the planned September 1 increase to $3 / $15 (official pricing docs, checked September 2, 2026).
  • Haiku 4.5 — fastest. 200K / 64K tokens. $1 / $5.

Note: pricing and specs were re-checked against the official Pricing and Models docs on September 2, 2026 and may change. Check the official docs for the latest.

Opus 5 — for hard work

Best for the most complex reasoning, long-horizon agentic — autonomous, multi-step task execution using tools like browsers or terminals — coding, and tasks needing high autonomy. Great when accuracy is paramount or you must handle a lot at once. In return it is the most expensive and overkill for light work.

Sonnet 5 — for most work

Released on June 30, 2026, Sonnet 5 is described by Anthropic as the most agentic Sonnet model yet. According to Anthropic's announcement, its performance in coding, reasoning, and tool use is a substantial improvement over Sonnet 4.6, and it is now the default model on Free and Pro plans. Its $2 / $10 pricing was announced as introductory through August 31, 2026, but has since been confirmed as the standard price (the planned increase to $3 / $15 was cancelled), so it now costs less than Sonnet 4.6 did. If you are unsure which model to use, start here.

Heads up: Sonnet 5 uses a new tokenizer — the engine that converts text into tokens — which produces approximately 30% more tokens for the same input compared to Sonnet 4.6. If you are using the API and have set token budgets or cost estimates based on Sonnet 4.6, recount them against Sonnet 5.

Haiku 4.5 — fast and lightweight

The fastest of the three. Good when quick responses matter, when you process simple, repetitive tasks in bulk, or when you need to save on cost. Note its context window (how much it reads at once) is the smallest at 200K tokens.

How to choose

What kind of task is it?Complex reasoningor accuracy critical→ Opus 5General work, codingor balanced performance→ Sonnet 5Simple, repetitive, orhigh-volume & low-cost→ Haiku 4.5Not sure? Start with Sonnet 5 → upgrade to Opus 5 if needed, or switch to Haiku 4.5 for speed/cost
  • Task is hard and accuracy matters → Opus 5
  • General work, a solid balance → Sonnet 5 (default on Free & Pro)
  • Speed, volume, cost savings → Haiku 4.5

In claude.ai chat you can switch models on screen; via the API — the interface developers use to call Claude from their own programs — you specify a model ID (e.g. claude-opus-5, claude-sonnet-5). Check the official docs for each model's exact ID. Note there is also a higher tier above these three (Fable 5.1, released September 1, 2026; the earlier Fable 5 is now legacy), but its availability terms differ from the standard models and need separate confirmation.

Related: Opus 4.8 New Features Guide · Opus 4.8 Complete Guide.

Was this helpful?

Keep reading

Readers read this next

Opus vs Sonnet: Opus 5.5 vs Sonnet 5 — Which Claude Model Should You Use?

Claude Opus 5 and Sonnet 5 differ in price, speed, context window, and thinking mode. Here's how to pick the right one in the API or the Claude.ai app.

435

Is Claude Fable 5.1 Nerfed? The Real Reasons Behind the Complaints

Some users feel Claude Fable 5.1 got weaker overnight. Here's what Anthropic's own documentation says is actually happening, and how to check your settings.

29

Back on Claude Fable? Here's What to Check First

If you tried another AI and came back to Claude Fable, here's how to switch to Fable 5.1, what changed in subscription terms, and how the automatic fallback to Opus works.

37

What Is the GPT-6 Astra 'Sweet Spot'? Choosing a Reasoning Effort Setting (and How Claude Compares)

GPT-6 Astra's reasoning effort setting has a widely discussed cost-to-performance 'sweet spot.' Here's what that means, and how the pricing compares to Claude Opus 5 and Sonnet 5.

67