claude-haiku-5-5 — API, Pricing & Context Window | Vivgrid

claude-haiku-5-5 is the latest model in Anthropic's Haiku tier and the fastest model in the current Claude lineup, built for high-volume, latency-sensitive work such as classification, routing, extraction, and subagent tasks. Succeeding claude-haiku-4-5, it moves to a 1M-token context window and 128K-token max output at a lower per-token price, and runs with adaptive thinking on by default.

On Vivgrid, claude-haiku-5-5 is geo-distributed across AMER, EMEA, and APAC and reachable on both the native Messages API (/messages) and Chat Completions (/chat/completions). It uses the same unified Vivgrid API key as the rest of the catalog, so moving an agent from claude-haiku-4-5 is a model-string change. It uses Anthropic's newer tokenizer, so the same text counts as roughly 30% more tokens than on claude-haiku-4-5.

Specifications

Pricing

Pricing in USD per 1M tokens.

What one request costs

The rates above applied to a 20K-token prompt with a 2K-token reply — a typical single agent turn, small enough that every request is billed at the base rate.

Prompt caching is billed at $0.125 per 1M cache-write tokens on the 5-minute TTL and $0.20 per 1M on the 1-hour TTL, with cached reads at $0.01.

Prompts over 100K tokens (uncached input, cache reads, and cache writes combined) are billed at the long-context rate: $0.50 input / $0.05 cached read / $0.625 5-minute cache write / $1.00 1-hour cache write / $2.50 output per 1M tokens.

Use claude-haiku-5-5 in your coding agent

claude-haiku-5-5 works with the same Vivgrid key and https://api.vivgrid.com/v1 endpoint as every other model in the catalog. Grab your API key from the Vivgrid Console, then drop one of the configurations below into your agent of choice.

Claude Code

Install Claude Code:

npm install -g @anthropic-ai/claude-code

claude-haiku-5-5 is served on Vivgrid's native Messages API (/messages), so Claude Code connects directly. Point it at Vivgrid and select the model in ~/.claude/settings.json:

{
  "env": {
    "ANTHROPIC_AUTH_TOKEN": "viv-xxxxxxxxxxxxx",
    "ANTHROPIC_BASE_URL": "https://api.vivgrid.com/",
    "ANTHROPIC_MODEL": "claude-haiku-5-5"
  }
}

Then start Claude Code in your project:

claude

Want different Vivgrid models for the Opus, Sonnet and Haiku slots? See the full Claude Code tutorial.

OpenCode

OpenCode has Vivgrid built in, so you can use claude-haiku-5-5 without writing any config: open /models, search for vivgrid, paste your API key and pick the model. The OpenCode tutorial walks through it step by step.

Prefer to configure it by hand? Install OpenCode:

npm i -g opencode-ai

Declare claude-haiku-5-5 under the Vivgrid provider in your global OpenCode config, ~/.config/opencode/opencode.json. The @ai-sdk/anthropic provider talks to Vivgrid over the Messages API (/v1/messages):

{
  "$schema": "https://opencode.ai/config.json",
  "model": "vivgrid/claude-haiku-5-5",
  "provider": {
    "vivgrid": {
      "npm": "@ai-sdk/anthropic",
      "name": "Vivgrid",
      "options": {
        "baseURL": "https://api.vivgrid.com/v1",
        "apiKey": "{env:VIVGRID_API_KEY}"
      },
      "models": {
        "claude-haiku-5-5": {
          "name": "claude-haiku-5-5",
          "tool_call": true,
          "attachment": true,
          "modalities": { "input": ["text", "image"], "output": ["text"] },
          "limit": { "context": 1000000, "output": 128000 }
        }
      }
    }
  }
}

Export your key and launch OpenCode:

export VIVGRID_API_KEY="viv-xxxxxxxxxxxxx"
opencode

Use /models to switch between claude-haiku-5-5 and other Vivgrid models mid-session. See the full OpenCode tutorial.

Pi

Install Pi:

npm install -g --ignore-scripts @earendil-works/pi-coding-agent

Add claude-haiku-5-5 to a Vivgrid provider in ~/.pi/agent/models.json. Pi talks to Vivgrid over the native Messages API (/v1/messages) with "api": "anthropic-messages". authHeader sends the key as a Bearer token. Keep it as its own provider, since a provider has a single base URL and this one differs from the Chat Completions models':

{
  "providers": {
    "vivgrid-anthropic": {
      "baseUrl": "https://api.vivgrid.com",
      "apiKey": "viv-xxxxxxxxxxxxx",
      "authHeader": true,
      "api": "anthropic-messages",
      "compat": {
        "sendSessionAffinityHeaders": true
      },
      "models": [
        {
          "id": "claude-haiku-5-5",
          "input": ["text", "image"],
          "contextWindow": 1000000,
          "maxTokens": 128000
        }
      ]
    }
  }
}

Make it the model Pi starts with in ~/.pi/agent/settings.json:

{
  "defaultProvider": "vivgrid-anthropic",
  "defaultModel": "claude-haiku-5-5"
}

Then verify and start Pi:

pi --list-models vivgrid-anthropic
pi

Or run a one-shot prompt without entering the TUI:

pi -p "explain the auth flow in this repo" --model vivgrid-anthropic/claude-haiku-5-5

See the full Pi tutorial.

API surfaces

The Vivgrid Console catalogs claude-haiku-5-5 on 2 surfaces, all behind the same API key. The endpoint shown is the one the Console previews for that surface.

Quick start

Get an API key from the Vivgrid Console, then call claude-haiku-5-5 directly.

curl https://api.vivgrid.com/v1/chat/completions \
  -H "Authorization: Bearer $VIVGRID_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-haiku-5-5",
    "messages": [
      { "role": "user", "content": "Say hello in English, Chinese and Spanish." }
    ],
    "stream": true
  }'

curl https://api.vivgrid.com/v1/messages \
  -H "Authorization: Bearer $VIVGRID_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-haiku-5-5",
    "max_tokens": 1024,
    "messages": [
      { "role": "user", "content": "Say hello in English, Chinese and Spanish." }
    ],
    "stream": true
  }'

Choosing within the Claude 5 family

Prices are USD per 1M tokens at the base rate. Rows marked this page are the model described here.

Ideal use cases

• High-volume classification, extraction, and routing
• Subagents and fast steps inside larger agent pipelines
• Real-time assistants where latency matters most
• Teams on claude-haiku-4-5 moving to a 1M-token context at a lower price

Related models

• claude-haiku-4-5 — the prior Haiku release
• claude-sonnet-5-5 — the latest Sonnet model, a step up in capability
• gpt-6-luna — the fast, low-cost gpt-6 model