Claude Opus 5

Who Should Read

A backend developer using the Claude API for coding agents or workflow automation who is weighing a model upgrade against cost and performance

Core Mechanics

• Opus 5 costs the same as Opus 4.8
  • It is priced at $5 per million input tokens and $25 per million output tokens, same as Opus 4.8.
  • The API model name is claude-opus-5.
• It nears Fable 5 at half the cost
  • On CursorBench 3.2 at max effort, it scores within 0.5% of Fable 5's peak at half the cost per task.
  • On Frontier-Bench v0.1, it beats all other models and more than doubles Opus 4.8 at lower cost per task.
• It leads on business automation and computer use
  • On Zapier AutomationBench, its pass rate is about 1.5x the next-best model at the same cost per task.
  • On OSWorld 2.0, it beats Fable 5's best result at just over a third of the cost.
  • On ARC-AGI 3, its score is 3x the next-best model.
• The effort setting trades quality for tokens
  • Effort lets you favor intelligence or save tokens for faster, cheaper results.
  • Even at its lowest effort on Zapier AutomationBench, it passes more tasks than any other model.

Community Reactions

How to Apply

• Swap the model name to claude-opus-5 and compare
  • Take 5 to 10 prompts you currently run on Opus 4.8 and change only the model name.
  • Record correctness and token usage side by side.
• Measure cost versus quality at each effort level
  • Run the same task set from the lowest effort up to max.
  • Set the lowest level that gives sufficient quality as your default.
• Use mid-conversation tool changes in agents
  • Find agents that add or remove tools mid-conversation.
  • Switch to the beta feature and check that the prompt cache stays valid.
• Enable automatic fallback for security-related requests
  • If you have security analysis features, log requests that hit the classifiers.
  • Turn on the automatic fallback beta and check that requests route to another model instead of being blocked.

Try It

# Run in Claude Code or a terminal. First: pip install anthropic, and set ANTHROPIC_API_KEY

import anthropic

client = anthropic.Anthropic()
prompt = "{{prompt to compare}}"

for model in ["claude-opus-4-8", "claude-opus-5"]:
    r = client.messages.create(
        model=model,
        max_tokens=2000,
        messages=[{"role": "user", "content": prompt}],
    )
    print(model, r.usage.input_tokens, r.usage.output_tokens)
    print(r.content[0].text[:500])

# Next: expand to 5-10 prompts and tabulate correctness and output tokens. Check the docs for the exact Opus 4.8 model ID and substitute it.

Terminology

Related Papers

• Mercury 3: Diffusion at scale

  Scaling diffusion LLM 40x gives Mercury 3: 4–8x faster than speed-tier models, similar quality

• Timestral: Mistral's time-series foundation models

  A 209M time-series model ranks first on three benchmarks and runs on CPU.

• DeepSeek V4.1 Flash 1017 (Max)

  DeepSeek V4.1 Flash 1017 Max scores 45 on the Intelligence Index at 211 tok/s and $0.30 input.

• If you don't run your own LLM proxy, how do you know whats going on?

  Running your own LLM proxy shows which keys, models, and cache misses drive your usage.

• DeepSeek: DeepSeek v4.1 Flash 1017

  DeepSeek V4.1 Flash 1017: $0.05/M input, 1M context, built for agents

• DeepSeek-v4.1-Flash-1017

  DeepSeek V4.1 Flash minor update lifts coding and agent benchmarks while cutting run time and output length

• Show HN: Llmspeed – feel how fast a local LLM runs before you buy the hardware

Related Resources

• https://www.anthropic.com/news/claude-opus-5
• https://news.ycombinator.com/item?id=49038433