LLM Model LandscapeFoundational

Constitutional AI Principles

The four Claude tiers, adaptive thinking and effort, the tool-use rules that differ from other providers, prompt caching, and where you can run it.

By Amit UpadhyayPublished 13 Oct 202611 min

Step 1 / 4

Speed

What's happening

Claude Model Family

claude-sonnet-5-5: best balance of speed/intelligence for most production tasks

💡 Analogy

Like a professional services firm — different tiers for different cost/quality tradeoffs.

Claude Sonnet 5.5 (September 2026) is Anthropic's production workhorse. Excellent for code, analysis, and agents. Haiku 5.5 is fast and cheap; Opus 5.5 and Fable 5.1 sit above it.

Console Output

> model: claude-sonnet-5-5
> Context: 1M tokens
> max_tokens: 1024
> System: code reviewer persona

●claude.py

Claude Model FamilyL5–8

1import anthropic

3client = anthropic.Anthropic() # uses ANTHROPIC_API_KEY env var

5# Claude Sonnet 5.5 — the production workhorse

6message = client.messages.create(

7 model="claude-sonnet-5-5",

8 max_tokens=1024,

9 system="You are a code reviewer. Be concise and specific.",

10 messages=[

11 {"role": "user", "content": "Review this Python function for bugs:\n\ndef divide(a, b):\n return a / b"}

12 ]

13)

14print(message.content[0].text)

15print(f"Input: {message.usage.input_tokens}, Output: {message.usage.output_tokens}")

17# Streaming for better UX

18with client.messages.stream(

19 model="claude-sonnet-5-5",

20 max_tokens=512,

21 messages=[{"role": "user", "content": "Write a sorting algorithm"}],

22) as stream:

23 for text in stream.text_stream:

24 print(text, end="", flush=True) # stream tokens in real time

26# Tool use (function calling)

27tools = [{

28 "name": "search_docs",

29 "description": "Search documentation for a query",

30 "input_schema": {

31 "type": "object",

32 "properties": {"query": {"type": "string"}},

33 "required": ["query"]

34 }

35}]

36response = client.messages.create(

37 model="claude-sonnet-5-5",

38 max_tokens=256,

39 tools=tools,

40 messages=[{"role": "user", "content": "Find docs about async/await"}],

41)

41 lines

Steps · 4

What's happening

Claude Model Family

claude-sonnet-5-5: best balance of speed/intelligence for most production tasks

💡 Analogy

Like a professional services firm — different tiers for different cost/quality tradeoffs.

Claude Sonnet 5.5 (September 2026) is Anthropic's production workhorse. Excellent for code, analysis, and agents. Haiku 5.5 is fast and cheap; Opus 5.5 and Fable 5.1 sit above it.

Console Output

> model: claude-sonnet-5-5
> Context: 1M tokens
> max_tokens: 1024
> System: code reviewer persona

Next: Test →

Claude Anthropic adaptive-thinking prompt-caching tool-use

← PreviousLlama: Open-Weight Models for Production Next →GPT: OpenAI's Model Family in Practice

Continue reading

All LLM Model Landscape articles →

LLM Model Landscape11 min

GPT: OpenAI's Model Family in Practice

The GPT-6 tiers, what reasoning effort replaced, when to use the Responses API over Chat Completions, how the bill is calculated, and what is different on Azure.

LLM Model Landscape11 min

Llama: Open-Weight Models for Production

The Llama lineup and licence, serving with Ollama and vLLM, GPU memory sizing, and an honest look at when a hosted API is still the better choice.

LLM Model Landscape10 min

Evals for Choosing a Model

Why public benchmarks do not transfer to your workload, how to build a model-selection eval, and why cost per successful task beats price per token.

Discussion

LLM Model LandscapeFoundational

Constitutional AI Principles

The four Claude tiers, adaptive thinking and effort, the tool-use rules that differ from other providers, prompt caching, and where you can run it.

By Amit UpadhyayPublished 13 Oct 202611 min

Step 1 / 4

Speed

What's happening

Claude Model Family

claude-sonnet-5-5: best balance of speed/intelligence for most production tasks

💡 Analogy

Like a professional services firm — different tiers for different cost/quality tradeoffs.

Claude Sonnet 5.5 (September 2026) is Anthropic's production workhorse. Excellent for code, analysis, and agents. Haiku 5.5 is fast and cheap; Opus 5.5 and Fable 5.1 sit above it.

Console Output

> model: claude-sonnet-5-5
> Context: 1M tokens
> max_tokens: 1024
> System: code reviewer persona

●claude.py

Claude Model FamilyL5–8

1import anthropic

3client = anthropic.Anthropic() # uses ANTHROPIC_API_KEY env var

5# Claude Sonnet 5.5 — the production workhorse

6message = client.messages.create(

7 model="claude-sonnet-5-5",

8 max_tokens=1024,

9 system="You are a code reviewer. Be concise and specific.",

10 messages=[

11 {"role": "user", "content": "Review this Python function for bugs:\n\ndef divide(a, b):\n return a / b"}

12 ]

13)

14print(message.content[0].text)

15print(f"Input: {message.usage.input_tokens}, Output: {message.usage.output_tokens}")

17# Streaming for better UX

18with client.messages.stream(

19 model="claude-sonnet-5-5",

20 max_tokens=512,

21 messages=[{"role": "user", "content": "Write a sorting algorithm"}],

22) as stream:

23 for text in stream.text_stream:

24 print(text, end="", flush=True) # stream tokens in real time

26# Tool use (function calling)

27tools = [{

28 "name": "search_docs",

29 "description": "Search documentation for a query",

30 "input_schema": {

31 "type": "object",

32 "properties": {"query": {"type": "string"}},

33 "required": ["query"]

34 }

35}]

36response = client.messages.create(

37 model="claude-sonnet-5-5",

38 max_tokens=256,

39 tools=tools,

40 messages=[{"role": "user", "content": "Find docs about async/await"}],

41)

41 lines

Steps · 4

What's happening

Claude Model Family

claude-sonnet-5-5: best balance of speed/intelligence for most production tasks

💡 Analogy

Like a professional services firm — different tiers for different cost/quality tradeoffs.

Claude Sonnet 5.5 (September 2026) is Anthropic's production workhorse. Excellent for code, analysis, and agents. Haiku 5.5 is fast and cheap; Opus 5.5 and Fable 5.1 sit above it.

Console Output

> model: claude-sonnet-5-5
> Context: 1M tokens
> max_tokens: 1024
> System: code reviewer persona

Next: Test →

Claude Anthropic adaptive-thinking prompt-caching tool-use

← PreviousLlama: Open-Weight Models for Production Next →GPT: OpenAI's Model Family in Practice

Continue reading

All LLM Model Landscape articles →

LLM Model Landscape11 min