Deploy Claude Opus 5 on Paperclip: Model IDs & Pricing

Anthropic’s current frontier lineup is Claude Opus 5, Sonnet 5, Haiku 4.5, and — for long-horizon agentic work — Fable 5.1. If you are running AI agents on Paperclip, every model in the lineup is a drop-in swap away: change the model ID in your agent config and the next run uses it. Model IDs and pricing in this guide were verified against Anthropic’s model documentation in September 2026.

Bottom line: Sonnet 5 costs $2 / $10 per million tokens (input/output) — permanent list pricing, not a promotion — and handles most production agent workloads. Opus 5, the flagship for complex agentic coding and enterprise work, is $5 / $25 with a 1M-token context. Swap the model ID in your Paperclip agent config — no code changes required.

The current Claude lineup at a glance

Model	Best for (Anthropic’s positioning)	Input / output per 1M tokens	Context
Claude Opus 5	Complex agentic coding and enterprise work	$5 / $25	1M tokens
Claude Sonnet 5	Best combination of speed and intelligence; most workloads	$2 / $10	1M tokens
Claude Haiku 4.5	Fastest model with near-frontier intelligence	$1 / $5	200K tokens
Claude Fable 5.1	Demanding reasoning and long-horizon agentic work	see Anthropic’s pricing	—

Anthropic’s own guidance: if you’re unsure which model to use, start with Claude Opus 5 for most workloads; use Fable 5.1 when your evals on Opus 5 at higher effort still fall short. In practice, most Paperclip agents default to Sonnet 5 for cost efficiency and reserve Opus 5 for the hardest coding and enterprise workloads.

If you’re coming from Opus 4.8 or earlier, you’re moving up a generation — Anthropic publishes a dedicated migrating-to-Opus-5 guide, and the switch on Paperclip is the same one-line change described below.

Why this matters if you run agents on Paperclip

1. Sonnet 5’s $2 / $10 pricing is permanent

When Sonnet 5 launched in June 2026, its $2 / $10 pricing carried an “introductory” label. That has expired into permanent list pricing — verified against Anthropic’s docs in September 2026. There is no reason left to budget against the old $3 / $15 figure.

2. Opus 5 raises the ceiling for coding agents

Opus 5 is Anthropic’s flagship for complex agentic coding and enterprise work, with a permanent 1M-token context. Long-horizon coding, large codebase refactors, and multi-hour autonomous runs fit in a single context window — no manual chunking.

3. Fable 5.1 covers the long-horizon gap

For agents doing demanding reasoning over long horizons — research sweeps, multi-day monitoring loops — Anthropic positions Fable 5.1 above Opus 5. If an agent stalls at Opus 5 even at higher effort, that’s the model to trial next.

How to upgrade your Paperclip agent

On HostAgentes managed Paperclip

1. Open your agent in the dashboard → Settings → Model
2. Select Claude Opus 5, Claude Sonnet 5, or another model from the dropdown
3. Save — no redeploy required. The next run uses the new model.

Your existing BYOK Anthropic API key, system prompts, and tool definitions continue to work unchanged.

In Paperclip YAML config

agent:
  name: "customer-support-agent"
  model:
    provider: anthropic
    id: claude-sonnet-5
    temperature: 0.3
    max_tokens: 8000
  tools:
    - name: search_knowledge_base
    - name: escalate_to_human

Use id: claude-opus-5 for agents that need the extra coding/reasoning depth.

Cost monitoring after the swap

Watch your Paperclip monitoring dashboard for the first 48 hours:

• Success rate should hold steady or rise on Sonnet 5 for tool-use and browsing agents
• Total spend should fall noticeably versus any Opus tier, even after re-measuring tokens per task
• Deep coding agents that regress on Sonnet 5 belong on Opus 5 — the flagship tier exists precisely for that workload

Which agents benefit most

1. General support, research, and knowledge-work agents — Sonnet 5 at $2 / $10 is the default; nothing in the lineup beats its price/performance for this shape.
2. Complex agentic coding and enterprise workloads — Opus 5, with its 1M-token context for whole-codebase work.
3. High-volume routing and classification — Haiku 4.5 at $1 / $5 keeps unit costs near zero.
4. Long-horizon reasoning agents — Fable 5.1, when Opus 5 at higher effort still falls short in your evals.

Frequently asked questions

What is the Claude Opus 5 model ID? claude-opus-5. Set it in your Paperclip agent config under model.id (with provider: anthropic), or pick Claude Opus 5 in the dashboard under Settings → Model.

How much does it cost to run agents on Opus 5 and Sonnet 5? Opus 5: $5 / $25 per million input/output tokens. Sonnet 5: $2 / $10 — permanent pricing, no longer an introductory rate. A typical agent run spends about 1,000 input and 100 output tokens, so 1M runs cost roughly $7.50 on Opus 5 or $3.00 on Sonnet 5 before batch and caching discounts.

Do I need to change my prompts when switching models? No. Prompts, system prompts, and tool definitions carry over. Sonnet 5 ships an updated tokenizer, so the same input can map to 1.0–1.35× more tokens than older Sonnet versions — re-measure tokens per task after the swap instead of assuming the sticker price scales linearly.

Should I use Sonnet 5 or Opus 5 for my agent? Start with Sonnet 5 — Anthropic positions it as the best combination of speed and intelligence for most workloads. Move to Opus 5 for complex agentic coding and enterprise work, and trial Fable 5.1 for long-horizon tasks where Opus 5 falls short. Run both on your own workload and let the success rate decide.

Can I use my own Anthropic API key on HostAgentes? Yes. BYOK is supported across all Paperclip and OpenClaw plans. Your Anthropic invoices stay on your Anthropic account; HostAgentes only charges for infrastructure.