Fakturert etter faktisk brukte tokens; ubrukt del av reservasjonen refunderes.
Ny her?
10 gratis credits (0,10 USD) når du registrerer deg med Google
Kan brukes 24 timer etter registrering, opptil 5 kjøringer per dag og maksimalt 2 credits per kjøring. Andre påloggingsmetoder starter uten credits. Nok for 16 kjøringer av denne modellen.
02
Om Claude Haiku 4.5
Kort sagtPer 23. september 2026
Claude Haiku 4.5 er en modell fra Anthropic i kategorien Multimodal. På Railwail koster Claude Haiku 4.5 1,20 USD per 1M inndatatokens og 6,00 USD per 1M utdatatokens. Kontekstvinduet inneholder 200 000 tokens, og ett svar kan være opptil 64 000 tokens langt.
Claude Haiku 4.5 brings Sonnet-class reliability to the cost-and-latency tier. 200K context, full vision input, native tool use and extended thinking on demand. Designed for high-volume agents, real-time chat, classification, simple coding subagents, and consumer applications where p50 latency matters more than peak reasoning.
Bakgrunn
Om Anthropic
Grunnlagt 2021 · San Francisco, USA
Anthropic was founded in 2021 by Dario and Daniela Amodei. The Claude model family launched in March 2023 with Haiku as the lightweight tier optimized for low latency and cost. Successive Haiku generations have steadily closed the gap to Sonnet: Haiku 3 (March 2024), Haiku 3.5 (October 2024), Haiku 4 (May 2025) and Haiku 4.5 (late 2025/early 2026). Anthropic is structured as a Public Benefit Corporation focused on AI safety, with total funding exceeding $20 billion and a 2026 valuation above $150 billion.
Claude Haiku 4.5 is the fastest and cheapest model in the Claude 4.x family. Architecturally it is a smaller decoder-only Transformer distilled from the Sonnet/Opus 4 training stack, optimized aggressively for inference latency and throughput while retaining the full Claude feature set (vision, extended thinking, tool use). It was trained on the same multi-trillion-token curated corpus as its larger siblings with a knowledge cutoff in early 2026. Post-training combined RLHF, Constitutional AI, RLAIF and distillation from larger teacher models. Despite its small size, Haiku 4.5 supports the same 200K context window and tool-use API as Sonnet 4.6 and Opus 4.7, making it a drop-in cheaper option for high-concurrency workloads.
Parametere
Undisclosed (estimated tens of billions of parameters)
Kontekst
200 000 tokens
Evner
Lowest p50 latency in the Claude 4.x family
200K token context window matching the larger Claude models
Vision input for images, PDFs and screenshots
Native tool use with parallel tool calling
Optional extended thinking mode for harder reasoning tasks
Sonnet-class reliability at one-third the cost
Strong instruction following and structured JSON output
Constitutional-AI alignment with low refusal rate
Available via Anthropic API, Bedrock and Vertex AI
Pretrained on the same curated multi-trillion-token mixture used by Sonnet and Opus 4.x. Post-training includes distillation from larger Anthropic models alongside RLHF, Constitutional AI and RLAIF. Knowledge cutoff in early 2026.
Lisens: Proprietary commercial license via Anthropic API, Amazon Bedrock and Google Vertex AI.
Sikkerhetstesting: Same Responsible Scaling Policy ASL-3 safeguards as Sonnet and Opus, with constitutional classifiers and third-party evaluations.
Kjente begrensninger
Below Sonnet 4.6 on hard reasoning and agentic coding benchmarks
No native audio or video input
Knowledge cutoff in early 2026
Less depth on niche or long-tail topics than larger models
curl https://railwail.com/api/v1/chat/completions \
-H "Authorization: Bearer $RAILWAIL_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "claude-haiku-4-5",
"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-haiku-4-5",
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-haiku-4-5",
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-haiku-4-5", [
{ role: "user", content: "Explain what a vector database is in two sentences." },
], { max_tokens: 1024 });
console.log(res.choices[0].message.content);
Claude Haiku 4.5 er en modell fra Anthropic i kategorien Multimodal. På Railwail kan du kalle den med en API-nøkkel gjennom Railwail API.
Hvor mye koster Claude Haiku 4.5 på Railwail?
På Railwail koster Claude Haiku 4.5 1,20 USD per 1M inndatatokens og 6,00 USD per 1M utdatatokens. Du betaler for det hver forespørsel faktisk bruker. Bruk betales fra forhåndskjøpte kreditter; 1 kreditt tilsvarer 0,01 USD.
Hvor stort er kontekstvinduet til Claude Haiku 4.5?
Kontekstvinduet til Claude Haiku 4.5 inneholder 200 000 tokens. Ett svar kan være opptil 64 000 tokens langt.
Hvor rask er Claude Haiku 4.5?
Det finnes ennå ikke nok målte kjøringer av Claude Haiku 4.5 på Railwail til å angi en kjøretid. Det avhenger av inndataene, innstillingene og belastningen hos leverandøren.
Er Claude Haiku 4.5 bedre enn BLIP?
Det avhenger av oppgaven. Claude Haiku 4.5 (Anthropic) og BLIP (Salesforce) er begge modeller i kategorien Multimodal. Sammenligningssiden viser prisene og spesifikasjonene deres side ved side.
Ja. Claude Haiku 4.5 godtar bilder som inndata i tillegg til tekst.
Hvordan bruker jeg Claude Haiku 4.5 gjennom API-en?
Opprett en Railwail API-nøkkel og send forespørselen din med modell-ID-en claude-haiku-4-5. Kodeeksempler for curl, Python og JavaScript finnes i API-delen av denne siden.
