Model guide

Claude Haiku 4.5: Where the Cheap, Fast Tier Actually Fits

Haiku 4.5 is the oldest model in the current Claude lineup and the only one you should never hand a feature. It is also, used correctly, the best money in the family. Here is the honest map of where it fits.

By Kylian Migot · Updated July 2026 · 6 min read

Quick answer

Claude Haiku 4.5 is Claude's fast/cheap tier: $1 in / $5 out per MTok, a 200k tokens window, released October 2025. It is built for subagent fan-outs, triage, summarization, and mechanical edits, not autonomous feature work. In Claude Code you select it with /model haiku.
What it is
The fast/cheap tier: subagents, triage, mechanical edits
Price
$1 in / $5 out per MTok (cheapest Claude tier)
Context
200k tokens (smallest in the family)
Not for
Autonomous feature work or complex multi-step coding
01

Where Haiku 4.5 Fits, Honestly

Claude Haiku 4.5 was released in October 2025, which makes it the oldest model in the current Claude lineup, everything above it, Sonnet 5, Opus 4.8, and Fable 5, shipped in mid-2026. It carries a 200k tokens window, the smallest in the family, and the API identifier is claude-haiku-4-5-20251001.

None of that is a reason to ignore it. At $1 in / $5 out per MTok it is the cheapest Claude by far, and there is a class of work, high-volume, bounded, mechanical, where paying more buys you nothing. The point of this page is to draw that boundary precisely: what Haiku is genuinely good at, and what it should never be asked to lead. For the full lineup context, see the best Claude model for coding.

02

Specs and Pricing

SpecClaude Haiku 4.5
VendorAnthropic
API identifierclaude-haiku-4-5-20251001
ReleasedOctober 2025
Price$1 in / $5 out per MTok
Context window200k tokens
TierFast (cheapest and oldest in the current lineup)
Claude Code/model haiku

Data verified July 18, 2026 against Anthropic's published pricing and docs. Anthropic has not published current agentic-coding benchmark scores for Haiku 4.5 that we can verify, and given its role that is the honest signal: this tier is not competing on SWE-bench, it is competing on price and latency.

03

The Cost Math vs Sonnet, Opus, and Fable

Every multiple below is computed from the verified list prices, not vibes:

ModelPriceInput vs HaikuOutput vs Haiku
Claude Haiku 4.5$1 in / $5 out per MTok1x1x
Claude Sonnet 5$3 in / $15 out per MTok3x3x
Claude Opus 4.8$5 in / $25 out per MTok5x5x
Claude Fable 5$10 in / $50 out per MTok10x10x

Data verified July 18, 2026. Two readings of this table. For a single task, the gap rarely matters, a story's tokens cost cents either way. For anything that runs wide or often, a fan-out of twenty subagents, a pipeline that triages every incoming issue, a summarizer that runs on every PR, the 5x-vs-Opus multiple is the difference between a rounding error and a line item. Note that Sonnet 5's intro pricing ($2 in / $10 out per MTok through August 31, 2026) temporarily narrows its gap to 2x on input, worth knowing while it lasts.

Ship your first agent today

Download AIDEN free and point it at your existing Claude Code or Codex setup. No credit card, running in minutes.

Download AIDEN free

Free to start · macOS 12+ · No credit card required

04

What Haiku Is Genuinely Good At

Subagent fan-outs

Low latency and low cost make wide parallel fan-outs affordable: searches, per-file checks, and bounded subtasks inside a larger workflow.

Triage & classification

Routing issues, labeling PRs, deciding which pipeline a request belongs in. High volume, clear criteria, no deep reasoning required.

Summarization

Condensing diffs, logs, threads, and session transcripts. The output is checked by whatever consumes it, so the cheap tier is the right tier.

Mechanical edits

Lint-fix style changes, rename sweeps, and template application, edits where the transformation is defined and the model just executes it.

Cost-capped pipelines

Anything that must run on every event with a hard budget. At $1 in / $5 out per MTok, always-on automation stays cheap enough to leave on.

Latency-sensitive steps

Interactive checks and quick lookups where a human is waiting and a flagship's extra seconds of thinking buy nothing.
05

What It's Not For

Haiku 4.5 is not built for complex autonomous coding, that is a design point, not a flaw. Multi-file feature work, debugging where the cause hides far from the symptom, and long unattended runs all demand reasoning depth this tier does not have, and the 200k tokens window rules out whole-repo context sessions that the 1M-window tiers handle natively.

06

The Pattern That Works: Opus Leads, Haiku Fans Out

The strongest Haiku deployment in 2026 is not Haiku alone, it is heterogeneous fleets. A lead session on Opus 4.8 owns the plan and the review, and fans bounded subtasks out to Haiku subagents: codebase searches, per-file summaries, mechanical change execution. Each subagent call costs 1/5th of the same call on Opus, so the fan-out can go wide while the judgment stays with the flagship.

The same logic applies one level up. On an AIDEN board, model routing is per story card: the mechanical cards, rename sweeps, doc syncs, lint burn-downs, run on Haiku, while Sonnet and Opus take the cards that need a real engineer's model. Cheap tier, cheap work; it is the whole trick.

FAQ

What is Claude Haiku 4.5 best used for?
Haiku 4.5 is Claude's fast/cheap tier, and it earns its keep on bounded, mechanical work: subagent fan-outs inside larger workflows, triage and classification, summarization, lint-fix style mechanical edits, and cost-capped pipelines. It is not built for complex autonomous coding, that is what Sonnet 5 and Opus 4.8 are for.
How much does Claude Haiku 4.5 cost?
Haiku 4.5 costs $1 in / $5 out per MTok, the cheapest Claude tier by far. Computed from list prices, Sonnet 5 costs 3x as much per token, Opus 4.8 costs 5x, and Fable 5 costs 10x, which is why Haiku is the default choice wherever volume, not capability, is the constraint.
Can Claude Haiku 4.5 handle feature work autonomously?
No, and pretending otherwise is the most common way teams sour on the fast tier. Haiku 4.5 is not built for complex autonomous coding: multi-file features, subtle debugging, and long unattended runs belong on Sonnet 5 at minimum, and on Opus 4.8 when reasoning is the bottleneck. Haiku's role is bounded tasks with clear success criteria.
What is Claude Haiku 4.5's context window?
200k tokens, the smallest window in the current Claude family, versus 1M tokens for Sonnet 5, Opus 4.8, and Fable 5. That is fine for the bounded tasks Haiku is suited to, but it rules out whole-repo sessions on large codebases.
How does Haiku work as a subagent model?
The strongest Haiku pattern in 2026 is heterogeneous fleets: a lead session on Opus 4.8 plans and reviews, then fans bounded subtasks out to Haiku subagents, searches, summaries, per-file mechanical changes. Each subagent costs 1/5th of an Opus call, so wide fan-outs stay affordable while the reasoning stays with the flagship.

Keep reading

Route the right model to the right story.

AIDEN runs your Claude Code and Codex CLIs on one kanban board, BYOK and local-first, so Haiku takes the mechanical cards while Opus and Sonnet carry the real work. Free for one project.

macOS 12+ · Bring your own Claude Code or Codex · Your code stays local