Log in
← All models

Claude Haiku 4.5

Budget

Anthropic's fast, cheap model for high-volume agent steps and subagents.

Anthropic
Maker
84/100
Overall
57
Value score
$2.20
Blended /1M
200K
Context
Proprietary
Weights
Pricing: $1.00 in · $5.00 out / 1M tokens · synced from OpenRouter · updated 2026-06-22

Scorecard

Output Quality8.2/10 · 73% community
Agentic Ability8.2/10 · 73% community
Speed9.0/10 · 77% community
Value for $8.3/10 · 74% community
Reliability9.3/10 · 79% community

Claude Haiku 4.5 is the small, fast member of the Claude family. It's a common pick for the many cheap calls an agent makes — file triage, summarization, tool routing, subagents — where Opus or Sonnet would be overkill. Quality is a clear step below the larger Claudes but the speed and price are the point.

Strengths
  • Very fast responses
  • Cheap enough for high-volume loops
  • Same Claude reliability and tool use
Trade-offs
  • Quality well below Sonnet/Opus on hard tasks
  • 200K context, not 1M

Agents that run Haiku 4.5 well

Cheap, fast subtasks and subagents inside a larger workflow.

Best model for Claude CodeBest model for ClineBest model for OpenClawBest model for HermesBest model for OpenCode

Claude Haiku 4.5 FAQ

Is Claude Haiku 4.5 good for AI agents?
Cheap, fast subtasks and subagents inside a larger workflow. On our editorial scoring it rates 8.2/10 for agentic ability and 8.2/10 for output quality. Very fast responses.
How much does Claude Haiku 4.5 cost?
Via OpenRouter, Claude Haiku 4.5 is priced at $1.00 in · $5.00 out / 1M tokens — a blended rate of about $2.20 per 1M tokens for typical input-heavy agent use.
What is Claude Haiku 4.5's context window?
Claude Haiku 4.5 has a 200K-token context window. Anthropic is the maker.
Can I self-host Claude Haiku 4.5?
No — Claude Haiku 4.5 is a proprietary, hosted-only model from Anthropic. You access it through an API (e.g. OpenRouter) rather than running the weights yourself.
What are the downsides of Claude Haiku 4.5?
The main trade-offs: quality well below sonnet/opus on hard tasks; 200k context, not 1m. It's strongest for cheap, fast subtasks and subagents inside a larger workflow.

Haiku 4.5 compared

Haiku 4.5 vs DeepSeek V4 FlashHaiku 4.5 vs GPT-5.4 MiniHaiku 4.5 vs Qwen3 235BHaiku 4.5 vs Flash LiteHaiku 4.5 vs Llama 4 MaverickHaiku 4.5 vs CodestralHaiku 4.5 vs Opus 4.8Haiku 4.5 vs Sonnet 4.6

The model is half the story — the agent is the other half

The model picks the moves; the agent runs the loop, the tools, and the guardrails. Once you've chosen a model, see which agent gets the most out of it.

Compare AI agents →