Claude Sonnet 4.6

Vendor: Anthropic Model ID: claude-sonnet-4-6 Capability: 200K context · tool use · vision · prompt caching · streaming Pricing: per-token, Sonnet tier (live rate) Sonnet 4.6 is what you should reach for first inside the 4-series. It keeps the 200K context window and tool-use shape of 4.5, while delivering measurably better structured output and tighter tool-call arguments. If you’re starting a new project, default here. If you’re on 4.5, the migration is a single string change — most production code sees a quality bump with no measurable latency or cost difference.

Request

curl https://llm.bytespike.ai/v1/messages \
  -H "x-api-key: $BYTESPIKE_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 4096,
    "messages": [
      {"role": "user", "content": "Extract the dates from this paragraph as JSON."}
    ]
  }'

Body parameters

Field	Type	Required	Default	Notes
`model`	string	yes	—	`claude-sonnet-4-6`
`messages`	array	yes	—	Conversation history.
`max_tokens`	integer	yes	—	Hard cap. Max for this model: 16384.
`system`	string \| array	no	—	Array form supports `cache_control`.
`temperature`	number	no	1.0	Range 0.0–1.0.
`top_p`	number	no	1.0	Nucleus sampling.
`tools`	array	no	—	Supported, including parallel tool calls.
`tool_choice`	object	no	`{"type":"auto"}`	`auto` / `any` / `tool` (named).
`stream`	boolean	no	false	SSE streaming.

Response

{
  "id": "msg_sonnet_…",
  "type": "message",
  "role": "assistant",
  "model": "claude-sonnet-4-6",
  "content": [
    {"type": "text", "text": "[\"2024-08-12\", \"2024-09-01\"]"}
  ],
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 87,
    "output_tokens": 24
  }
}

Code examples

curl https://llm.bytespike.ai/v1/messages \
  -H "x-api-key: $BYTESPIKE_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "max_tokens": 4096,
    "messages": [{"role": "user", "content": "Extract the dates from this paragraph as JSON."}]
  }'

import anthropic

client = anthropic.Anthropic(
    base_url="https://llm.bytespike.ai/v1",
    api_key="$BYTESPIKE_API_KEY",
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=4096,
    messages=[
        {"role": "user", "content": "Extract the dates from this paragraph as JSON."}
    ],
)

print(message.content[0].text)

import Anthropic from "@anthropic-ai/sdk"

const client = new Anthropic({
  baseURL: "https://llm.bytespike.ai/v1",
  apiKey: process.env.BYTESPIKE_API_KEY,
})

const message = await client.messages.create({
  model: "claude-sonnet-4-6",
  max_tokens: 4096,
  messages: [
    { role: "user", content: "Extract the dates from this paragraph as JSON." },
  ],
})

console.log(message.content[0].text)

Streaming

Set "stream": true for SSE in the standard Anthropic format. Estimated credits ship in the HTTP headers before the first event.

Cache control

cache_control blocks reduce cost on repeated prompts. Cache reads at the discounted rate visible in the pricing table. With Sonnet 4.6’s tighter tool-arg generation, retrieval-augmented agent loops see the biggest cache wins (system prompt + tool schema kept stable, only the user turn changes).

{
  "model": "claude-sonnet-4-6",
  "system": [
    {
      "type": "text",
      "text": "<long static system prompt>",
      "cache_control": {"type": "ephemeral"}
    }
  ],
  "messages": [...]
}

Errors

Code	Trigger	Billed?
400	Body validation failed	No
401	Missing / revoked key	No
402	Wallet exhausted	No
403	Scope denied / IP not allowlisted	No
429	Rate-limited	No
5xx	Upstream provider issue	No (auto-retry envelope)

When to use

Default Anthropic mid-tier — start here, benchmark against Opus / Haiku later.
Code generation / refactoring / structured extraction where Haiku is too imprecise.
Tool-heavy agents where parallel tool calls and tight argument JSON matter.
For higher throughput at lower quality, see Haiku 4.5.
For deeper reasoning across long contexts, see Opus 4.7.

Limits

Limit	Value
Context window	200K tokens
Max output	16384 tokens
Supports tool use	Yes (parallel)
Supports vision	Yes
Supports streaming	Yes
Supports prompt caching	Yes

API Reference

Protocol endpoints

OpenAI

Anthropic

DeepSeek

Moonshot

Zhipu

MiniMax

Image

Video

Account & billing

Utility

Request

Body parameters

Response

Code examples

Streaming

Cache control

Errors

When to use

Limits

​Request

​Body parameters

​Response

​Code examples

​Streaming

​Cache control

​Errors

​When to use

​Limits

Request

Body parameters

Response

Code examples

Streaming

Cache control

Errors

When to use

Limits