Claude Sonnet 5 Is Live: Near Opus 4.8, Priced Same as Official

​

Key Takeaways

• Strongest Sonnet yet: Anthropic released Claude Sonnet 5 on June 30, calling it “the most agentic Sonnet model yet,” with performance close to flagship Opus 4.8
• Big coding leap: 85.2% on SWE-bench Verified, roughly +5.6pp over Sonnet 4.6’s 79.6%; 80.4% on Terminal-Bench 2.1
• Priced same as official: APIYI now offers claude-sonnet-5 at exactly Anthropic’s official price (intro $2/$10, then $3/$15 per million tokens after Aug 31)
• Pure relay + high cache hit: Runs on pure AWS Bedrock Claude official-relay resources with a high cache-hit rate, keeping long agent chains cost-controlled
• A workhorse agent model: Plans, uses tools like browsers and terminals, and runs autonomously for long stretches — a high-value pick for agents

​

Background

​

Deep Dive

​

Core Features

​

Performance Highlights

• SWE-bench Verified rises from Sonnet 4.6’s 79.6% to 85.2%, more reliable on real repo-level fixes.
• 80.4% on Terminal-Bench 2.1 — strong for CLI/terminal agents, repo maintenance, and migration tools.

• Better at driving long chains to completion instead of stalling to ask for permission.
• Early users report it’s “almost as good as Opus 4.8, but faster and cheaper.”

​

Technical Specs

​

Real-World Use

​

Recommended Scenarios

1. Daily workhorse coding: repo-level fixes, refactors, bug hunting at Sonnet pricing with near-flagship results
2. Long agent chains: browser/terminal tool use, autonomous research and coding agents
3. CLI / repo-maintenance automation: strong Terminal-Bench results suit migration tools and batch scripts
4. Cost-sensitive high-concurrency calls: high cache-hit + Sonnet pricing for scaled deployments
5. Knowledge work: broad gains in reasoning, tool use, and writing

​

Code Examples

​

OpenAI format

from openai import OpenAI

client = OpenAI(
    api_key="your-apiyi-key",
    base_url="https://api.apiyi.com/v1"
)

response = client.chat.completions.create(
    model="claude-sonnet-5",
    messages=[
        {"role": "user", "content": "Review this code and flag potential bugs and improvements."}
    ]
)

print(response.choices[0].message.content)

​

Anthropic-native format

import anthropic

client = anthropic.Anthropic(
    api_key="your-apiyi-key",
    base_url="https://api.apiyi.com"
)

message = client.messages.create(
    model="claude-sonnet-5",
    max_tokens=8192,
    messages=[
        {"role": "user", "content": "Walk through the root cause of this production bug step by step and propose a fix."}
    ]
)

print(message.content[0].text)

​

In Claude Code

{
  "model": "claude-sonnet-5",
  "apiKey": "your-apiyi-key",
  "baseURL": "https://api.apiyi.com"
}

​

Best Practices

1. Daily workhorse: swap claude-sonnet-5 in for Sonnet 4.6 — better results, same price.
2. Leverage caching: put stable system prompts and code context up front to maximize cache-hit and cut cost.
3. Let long tasks run: allow longer autonomous runs in agent scenarios, reducing manual intervention.
4. Escalate when needed: switch to claude-opus-4-8 for the most complex architecture-level decisions.

​

Pricing & Availability

​

Pricing

​

Stack with Top-up Promotions

​

Where to Get It

• Site: apiyi.com
• OpenAI format: https://api.apiyi.com/v1
• Anthropic-native format: https://api.apiyi.com
• Model name: claude-sonnet-5

​

Summary & Recommendation

• Stronger: 85.2% SWE-bench Verified, performance near Opus 4.8
• Cheaper: Sonnet-tier pricing + high cache-hit
• Steadier: pure AWS Bedrock official-relay resources
• No markup: identical to Anthropic’s official pricing