Anthropic Messages

Bruk Anthropic Messages API gjennom CometAPI for å få tilgang til Claude-modeller med extended thinking, prompt caching, verktøybruk, web search/fetch, Streaming og effort control.

POST /v1/messages

Python

Basic

import anthropic

client = anthropic.Anthropic(
    base_url="https://api.cometapi.com",
    api_key="<COMETAPI_KEY>",
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system="You are a helpful assistant.",
    messages=[
        {"role": "user", "content": "Hello, world"}
    ],
)

print(message.content[0].text)

200

{
  "id": "<string>",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "<string>",
      "thinking": "<string>",
      "signature": "<string>",
      "id": "<string>",
      "name": "<string>",
      "input": {}
    }
  ],
  "model": "<string>",
  "stop_reason": "end_turn",
  "stop_sequence": "<string>",
  "usage": {
    "input_tokens": 123,
    "output_tokens": 123,
    "cache_creation_input_tokens": 123,
    "cache_read_input_tokens": 123,
    "cache_creation": {
      "ephemeral_5m_input_tokens": 123,
      "ephemeral_1h_input_tokens": 123
    }
  }
}

​

Oversikt

CometAPI støtter Anthropic Messages API direkte, og gir deg direkte tilgang til Claude-modeller med alle Anthropic-spesifikke funksjoner. Bruk dette endepunktet for Claude-eksklusive muligheter som extended thinking, prompt caching og effort control.

​

Kom raskt i gang

Bruk den offisielle Anthropic SDK-en — bare sett base-URL-en til CometAPI:

Python

import anthropic

client = anthropic.Anthropic(
    base_url="https://api.cometapi.com",
    api_key="<COMETAPI_KEY>",
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}],
)
print(message.content[0].text)

Både x-api-key og Authorization: Bearer-headere støttes for autentisering. De offisielle Anthropic SDK-ene bruker x-api-key som standard.

​

Extended Thinking

Aktiver Claudes trinnvise resonnement med parameteren thinking. Svaret inkluderer thinking-innholdsblokker som viser Claudes interne resonnement før det endelige svaret.

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=16000,
    thinking={
        "type": "enabled",
        "budget_tokens": 10000,
    },
    messages=[
        {"role": "user", "content": "Prove that there are infinitely many primes."}
    ],
)

for block in message.content:
    if block.type == "thinking":
        print(f"Thinking: {block.thinking[:200]}...")
    elif block.type == "text":
        print(f"Answer: {block.text}")

Thinking krever minst budget_tokens på 1,024. Thinking-tokens teller mot grensen for max_tokens — sett max_tokens høyt nok til å dekke både thinking og svaret.

---

​

Prompt Caching

Bufre store system-prompts eller samtaleprefikser for å redusere latenstid og kostnad ved påfølgende forespørsler. Legg til cache_control i innholdsblokker som skal bufres:

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system=[
        {
            "type": "text",
            "text": "You are an expert code reviewer. [Long detailed instructions...]",
            "cache_control": {"type": "ephemeral"},
        }
    ],
    messages=[{"role": "user", "content": "Review this code..."}],
)

Bruk av cache rapporteres i responsens usage-felt:

• cache_creation_input_tokens — tokens skrevet til cache (faktureres til en høyere sats)
• cache_read_input_tokens — tokens lest fra cache (faktureres til en redusert sats)

Prompt caching krever minst 1,024 tokens i den bufrede innholdsblokken. Innhold som er kortere enn dette, blir ikke bufret.

---

​

Streaming

Stream svar ved hjelp av Server-Sent Events (SSE) ved å sette stream: true. Hendelser kommer i denne rekkefølgen:

1. message_start — inneholder meldingsmetadata og innledende bruk
2. content_block_start — markerer starten på hver content-blokk
3. content_block_delta — inkrementelle tekstbiter (text_delta)
4. content_block_stop — markerer slutten på hver content-blokk
5. message_delta — endelig stop_reason og fullstendig usage
6. message_stop — signaliserer slutten på strømmen

with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=256,
    messages=[{"role": "user", "content": "Hello"}],
) as stream:
    for text in stream.text_stream:
        print(text, end="")

---

​

Kontroll av innsats

Kontroller hvor mye innsats Claude legger i å generere et svar med output_config.effort:

message = client.messages.create(
    model="claude-opus-4-6",
    max_tokens=4096,
    messages=[
        {"role": "user", "content": "Summarize this briefly."}
    ],
    output_config={"effort": "low"},  # "low", "medium", or "high"
)

---

​

Serververktøy

Claude støtter verktøy på serversiden som kjører på Anthropics infrastruktur:

Web Fetch

Hent og analyser innhold fra URL-er:

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Analyze the content at https://arxiv.org/abs/1512.03385"}
    ],
    tools=[
        {"type": "web_fetch_20250910", "name": "web_fetch", "max_uses": 5}
    ],
)

Web Search

Søk på nettet etter sanntidsinformasjon:

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "What are the latest developments in AI?"}
    ],
    tools=[
        {"type": "web_search_20250305", "name": "web_search", "max_uses": 5}
    ],
)

---

​

Eksempel på svar

Et typisk svar fra CometAPIs Anthropic-endepunkt:

{
  "id": "msg_bdrk_01UjHdmSztrL7QYYm7CKBDFB",
  "type": "message",
  "role": "assistant",
  "content": [
    {
      "type": "text",
      "text": "Hello!"
    }
  ],
  "model": "claude-sonnet-4-6",
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "usage": {
    "input_tokens": 19,
    "cache_creation_input_tokens": 0,
    "cache_read_input_tokens": 0,
    "cache_creation": {
      "ephemeral_5m_input_tokens": 0,
      "ephemeral_1h_input_tokens": 0
    },
    "output_tokens": 4
  }
}

---

​

Viktige forskjeller fra OpenAI-kompatibelt endepunkt

Funksjon	Anthropic Messages (/v1/messages)	OpenAI-Compatible (/v1/chat/completions)
Utvidet tenking	thinking-parameter med budget_tokens	Ikke tilgjengelig
Prompt caching	cache_control på content-blokker	Ikke tilgjengelig
Kontroll av innsats	output_config.effort	Ikke tilgjengelig
Web fetch/søk	Serververktøy (web_fetch, web_search)	Ikke tilgjengelig
Auth-header	x-api-key eller Bearer	Kun Bearer
Svarformat	Anthropic-format (content-blokker)	OpenAI-format (choices, message)
Modeller	Kun Claude	Flere tilbydere (GPT, Claude, Gemini, osv.)

Autorisasjoner

x-api-key · string · header · påkrevd

Your CometAPI key passed via the x-api-key header. Authorization: Bearer <key> is also supported.

Overskrifter

anthropic-version · string · standard: 2023-06-01

The Anthropic API version to use. Defaults to 2023-06-01.

Eksempel:

"2023-06-01"

anthropic-beta · string

Comma-separated list of beta features to enable. Examples: max-tokens-3-5-sonnet-2024-07-15, pdfs-2024-09-25, output-128k-2025-02-19.

Kropp

application/json

model · string · påkrevd

The Claude model to use. See the Models page for current Claude model IDs.

Eksempel:

"claude-sonnet-4-6"

messages · object[] · påkrevd

The conversation messages. Must alternate between user and assistant roles. Each message's content can be a string or an array of content blocks (text, image, document, tool_use, tool_result). There is a limit of 100,000 messages per request.

Show child attributes

max_tokens · integer · påkrevd

The maximum number of tokens to generate. The model may stop before reaching this limit. When using thinking, the thinking tokens count towards this limit.

Nødvendig område: x >= 1

Eksempel:

1024

system · string object[]

System prompt providing context and instructions to Claude. Can be a plain string or an array of content blocks (useful for prompt caching).

temperature · number · standard: 1

Controls randomness in the response. Range: 0.0–1.0. Use lower values for analytical tasks and higher values for creative tasks. Defaults to 1.0.

Nødvendig område: 0 <= x <= 1

top_p · number

Nucleus sampling threshold. Only tokens with cumulative probability up to this value are considered. Range: 0.0–1.0. Use either temperature or top_p, not both.

Nødvendig område: 0 <= x <= 1

top_k · integer

Only sample from the top K most probable tokens. Recommended for advanced use cases only.

Nødvendig område: x >= 0

stream · boolean · standard: false

If true, stream the response incrementally using Server-Sent Events (SSE). Events include message_start, content_block_start, content_block_delta, content_block_stop, message_delta, and message_stop.

stop_sequences · string[]

Custom strings that cause the model to stop generating when encountered. The stop sequence is not included in the response.

thinking · object

Enable extended thinking — Claude's step-by-step reasoning process. When enabled, the response includes thinking content blocks before the answer. Requires a minimum budget_tokens of 1,024.

Show child attributes

tools · object[]

Tools the model may use. Supports client-defined functions, web search (web_search_20250305), web fetch (web_fetch_20250910), code execution (code_execution_20250522), and more.

Show child attributes

tool_choice · object

Controls how the model uses tools.

Show child attributes

metadata · object

Request metadata for tracking and analytics.

Show child attributes

output_config · object

Configuration for output behavior.

Show child attributes

service_tier · enum<string>

The service tier to use. auto tries priority capacity first, standard_only uses only standard capacity.

Tilgjengelige alternativer: auto, standard_only

Svar

200 - application/json

Successful response. When stream is true, the response is a stream of SSE events.

id · string

Unique identifier for this message (e.g., msg_01XFDUDYJgAACzvnptvVoYEL).

type · enum<string>

Always message.

Tilgjengelige alternativer: message

role · enum<string>

Always assistant.

Tilgjengelige alternativer: assistant

content · object[]

The response content blocks. May include text, thinking, tool_use, and other block types.

Show child attributes

model · string

The specific model version that generated this response (e.g., claude-sonnet-4-6).

stop_reason · enum<string>

Why the model stopped generating.

Tilgjengelige alternativer: end_turn, max_tokens, stop_sequence, tool_use, pause_turn

stop_sequence · string | null

The stop sequence that caused the model to stop, if applicable.

usage · object

Token usage statistics.

Show child attributes