Claude Sonnet 4.6

Name: Claude Sonnet 4.6
Brand: Anthropic
SKU: claude-sonnet-4-6
Price: 3.6 USD
Availability: InStock

MultimodalÆldre versionTilgængelig

af Anthropicclaude-sonnet-4-6

Anthropic's balanced mid-tier model from February 2026. Best price/performance for production workloads: 5x cheaper than Opus, near-flagship quality.

Prøv det Få API-nøgle Sammenlign

Pris · 1M ind/ud: 3,60 US$ / 18,00 US$
Kontekst: 1.000.000 tokens
Maks. output: 128.000 tokens
Input → output: Tekst + Billede → Tekst
Udvikler: Anthropic
Opdateret: 23. september 2026

Dette er en ældre version af modellen.

Nyere version tilgængelig: Claude Sonnet 5

Playground

Prøv Claude Sonnet 4.6

Chat

3,60 US$/1M in

Send en besked. Svaret ankommer fuldt ud, når modellen er færdig (uden streaming).

Systemprompt

Maks. svarslængde (tokens)

TemperaturModelstandard

⌘/Ctrl + Enter

Denne kørsel

højst 0,0185 US$ · 1,85 credits reserveret

Fakturering efter faktisk brugte tokens; den ubrugte del af reservationen refunderes.

Ny her?

10 gratis credits (0,10 US$) når du tilmelder dig med Google

Kan bruges 24 timer efter tilmelding, op til 5 kørsler pr. dag og højst 2 credits pr. kørsel. Andre login-metoder starter uden credits. Nok til 5 kørsler af denne model.

Om Claude Sonnet 4.6

Kort sagtFra 23. september 2026

Claude Sonnet 4.6 er en model af Anthropic i kategorien Multimodal. På Railwail koster Claude Sonnet 4.6 3,60 US$ pr. 1M inputtokens og 18,00 US$ pr. 1M outputtokens. Kontekstvinduet indeholder 1.000.000 tokens, og et svar kan være op til 128.000 tokens langt. Nyere version: Claude Sonnet 5.

Released February 17, 2026, Claude Sonnet 4.6 is the workhorse of the 4.x Claude family. Delivers near-Opus quality on most coding, writing and analysis tasks at one-fifth the cost. 1M-token context window, extended thinking, full vision support, and the same tool-use loop as Opus 4.7. Suited to production agents, developer copilots, customer support, and most enterprise API scenarios.

Baggrund

Om Anthropic

Grundlagt 2021 · San Francisco, USA

Anthropic was founded in 2021 by Dario and Daniela Amodei together with senior researchers from OpenAI. The company is a Public Benefit Corporation focused on AI safety. Key research milestones include Constitutional AI (2022), Sparse Autoencoders for interpretability, and the Claude model family launched in March 2023. Claude Sonnet has been Anthropic's flagship 'balanced' tier since the Claude 3 family, with major releases including Sonnet 3.5 (June 2024), Sonnet 3.7 with extended thinking (February 2025), Sonnet 4 (May 2025), Sonnet 4.5 (late 2025) and Sonnet 4.6 (February 17, 2026). Investors include Google, Amazon, Spark Capital and Lightspeed; total funding exceeds $20 billion with a 2026 valuation above $150 billion.

Besøg Anthropic

Arkitektur

Decoder-only Transformer (balanced production model with extended thinking)

Claude Sonnet 4.6 was released February 17, 2026 as the mid-tier workhorse of the Claude 4.x family. It is a decoder-only Transformer engineered to deliver near-Opus quality at one-fifth the cost. Pretraining used the same multi-trillion-token mixture as Opus 4.7 with a knowledge cutoff in early 2026. Post-training combined RLHF, Constitutional AI, RLAIF, and reinforcement learning against verifiable rewards on coding, math and tool-use trajectories. Sonnet 4.6 supports the full Opus feature set: extended thinking mode, native vision, computer use, parallel tool calling, and 200K context. The training emphasis was on efficiency: Sonnet 4.6 outperforms Opus 4 on most coding and writing benchmarks while running ~2x faster and at one-fifth the cost. Anthropic positions Sonnet 4.6 as the recommended default for production workloads.

Parametre: Undisclosed (estimated dense model in the 100B-range)
Kontekst: 200.000 tokens

Funktioner

Near-Opus quality on most coding, writing and analysis tasks at 5x lower cost
Extended thinking mode with adjustable thinking budget
200K token context window with strong needle-in-haystack recall
Multimodal input: text, images, PDFs, charts and diagrams
Native computer use: screenshots, mouse and keyboard control
Parallel tool calling and stable multi-turn agent loops
Strong instruction following including XML tags, schemas and strict JSON
Constitutional-AI alignment with low refusal rate on benign questions
Multilingual fluency across 20+ languages
~2x faster latency than Opus 4.7 on equivalent tasks
Best for: production agents, developer copilots, customer support, most enterprise API scenarios.

Træning og licens

Pretrained on a curated multi-trillion-token mixture identical to the Opus 4.7 corpus, with heavy quality filtering, deduplication and copyright screening. Knowledge cutoff in early 2026. Post-training uses RLHF, Constitutional AI, RLAIF and RL against verifiable rewards.

Licens: Proprietary commercial license via Anthropic API, Amazon Bedrock and Google Vertex AI. Commercial use permitted under Anthropic's Usage Policy.

Sikkerhedstests: Deployed under Responsible Scaling Policy ASL-3 safeguards with constitutional classifiers, automated red-teaming, third-party evaluations and a public Bug Bounty programme.

Kendte begrænsninger

Below Opus 4.7 on the hardest agentic and reasoning benchmarks
No native audio or video input
Knowledge cutoff in early 2026; needs tool use for fresh information
Will refuse some legitimate red-team and security research prompts
No fine-tuning available to external customers

Forskningsartikler

Priser

Priser i US-dollar. Forbrug debiteres fra forudbetalte credits.

Input	3,60 US$ / 1M tokens
Output	18,00 US$ / 1M tokens

Faktureres efter de tokens, som hver anmodning faktisk bruger.
1 kredit = 0,01 US$

Omkostningsberegner

Prisberegner

Input-tokens

/ anmodning

Output-tokens

/ anmodning

Anmodninger

I alt

1,26 US$

126 credits

Pr. anmodning

0,0126 US$ · 1,26 credits

Hver anmodning rundes op til 0,01 credits.

API

Kald Claude Sonnet 4.6 med din Railwail API-nøgle. Brug dette model-ID i anmodningen:

claude-sonnet-4-6API-dokumentation Få en API-nøgle

curl https://railwail.com/api/v1/chat/completions \
  -H "Authorization: Bearer $RAILWAIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "messages": [
      {
        "role": "user",
        "content": "Explain what a vector database is in two sentences."
      }
    ],
    "max_tokens": 1024
  }'

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["RAILWAIL_API_KEY"],
    base_url="https://railwail.com/api/v1",
)

completion = client.chat.completions.create(
    model="claude-sonnet-4-6",
    messages=[
        {
            "role": "user",
            "content": "Explain what a vector database is in two sentences.",
        },
    ],
    max_tokens=1024,
)

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

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.RAILWAIL_API_KEY,
  baseURL: "https://railwail.com/api/v1",
});

const completion = await client.chat.completions.create({
  model: "claude-sonnet-4-6",
  messages: [
    {
      role: "user",
      content: "Explain what a vector database is in two sentences."
    }
  ],
  max_tokens: 1024
});

console.log(completion.choices[0].message.content);

// npm install railwail
import railwail from "railwail";

const rw = railwail(process.env.RAILWAIL_API_KEY);

const res = await rw.chat("claude-sonnet-4-6", [
  { role: "user", content: "Explain what a vector database is in two sentences." },
], { max_tokens: 1024 });

console.log(res.choices[0].message.content);

Indstil din nøgle som RAILWAIL_API_KEYOpret API-nøgle

Specifikationer

Model-ID: claude-sonnet-4-6
Udvikler: Anthropic
Kategori: Multimodal
Input: Tekst, Billede
Output: Tekst
Kontekstvindue: 1.000.000 tokens
Maks. output: 128.000 tokens
Fakturering: Efter forbrug (tokens eller GPU-tid)
Udgivet: 17. februar 2026
Livscyklus: Ældre version
Modelstørrelse: Undisclosed (estimated dense model in the 100B-range)
Licens: Proprietary commercial license via Anthropic API, Amazon Bedrock and Google Vertex AI. Commercial use permitted under Anthropic's Usage Policy.
Katalogelement opdateret: 23. september 2026

Inputparametre

Inputs og indstillinger fra modellens inputskema. Eksemplet i API-afsnittet viser, hvilke af dem API'en accepterer.

ParameterTypeStandardTilladte værdier

promptpåkrævet
User message
Type: Tekst
Standard: –
Tilladte værdier: op til 32.000 tegn
top_p
Type: Tal
Standard: 1
Tilladte værdier: 0 til 1
stream
Type: Ja/Nej
Standard: false
Tilladte værdier: –
image_url
Optional image URL to analyze
Type: Tekst
Standard: –
Tilladte værdier: –
max_tokens
Type: Heltal
Standard: 4096
Tilladte værdier: 1 til 32.000
temperature
Type: Tal
Standard: 1
Tilladte værdier: 0 til 1
system_prompt
Optional system instruction
Type: Tekst
Standard: –
Tilladte værdier: op til 8.000 tegn

Anvendelsestilfælde

Hvad det bruges til

Production coding assistants and IDE copilots
Customer support chatbots and triage agents
Enterprise document analysis and summarisation
Multi-turn agent loops with tool use
Computer-use automation
Content generation and editing pipelines

Alle modeller i Multimodal

Ofte stillede spørgsmål

Hvad er Claude Sonnet 4.6?

Claude Sonnet 4.6 er en model fra Anthropic i kategorien Multimodal. På Railwail kan du kalde den med en API-nøgle via Railwail API.

Hvad koster Claude Sonnet 4.6 på Railwail?

På Railwail koster Claude Sonnet 4.6 3,60 US$ pr. 1M inputtokens og 18,00 US$ pr. 1M outputtokens. Du betaler for det, som hver anmodning faktisk bruger. Forbrug betales fra forudbetalte credits; 1 credit svarer til 0,01 US$.

Hvad er kontekstvinduet for Claude Sonnet 4.6?

Kontekstvinduet for Claude Sonnet 4.6 indeholder 1.000.000 tokens. Et svar kan være op til 128.000 tokens langt.

Hvor hurtig er Claude Sonnet 4.6?

Der er endnu ikke nok målte kørsler af Claude Sonnet 4.6 på Railwail til at angive en udførelsestid. Det afhænger af inputtet, indstillingerne og belastningen hos provideren.

Er Claude Sonnet 4.6 bedre end Claude Sonnet 5?

Det afhænger af opgaven. Claude Sonnet 4.6 (Anthropic) og Claude Sonnet 5 (Anthropic) er begge modeller i kategorien Multimodal. Sammenligningssiden viser deres priser og specifikationer side om side.

Sammenlign Claude Sonnet 4.6 og Claude Sonnet 5

Kan Claude Sonnet 4.6 behandle billeder?

Ja. Claude Sonnet 4.6 accepterer billeder som input ud over tekst.

Hvordan bruger jeg Claude Sonnet 4.6 via API'en?

Opret en Railwail API-nøgle og send din anmodning med model-ID'en claude-sonnet-4-6. Kodeeksempler til curl, Python og JavaScript findes i API-sektionen på denne side.

Sammenlignelige modeller

Alle i denne kategori

Claude Opus 4.7Anthropic
Anthropic's April 2026 flagship. 87.6% on SWE-bench Verified, 3x higher image resolution, output self-verification, vision + reasoning.
6,00 US$/1M in
67 % dyrere pr. enhed
Sammenlign Claude Sonnet 4.6 og Claude Opus 4.7
Gemini 3 FlashGoogle DeepMind
Google's April 2026 fast multimodal model. Combines Gemini 3 Pro's reasoning with Flash-tier latency and price. Default model in the Gemini app.
0,60 US$/1M in
83 % billigere pr. enhed
Sammenlign Claude Sonnet 4.6 og Gemini 3 Flash
Gemini 3.1 ProGoogle DeepMind
Google DeepMind's February 2026 flagship. 1M-token (1,048,576) context, native multimodal (text/image/audio/video), Deep Think reasoning.
2,40 US$/1M in
33 % billigere pr. enhed
Sammenlign Claude Sonnet 4.6 og Gemini 3.1 Pro

Brug Claude Sonnet 4.6 via API'en

En API-nøgle til alle modeller på Railwail. Forbrug debiteres fra forudbetalte credits, 1 credit = 0,01 US$.

Få API-nøgle Se alle modeller