claude-mythos-5
Input price$10
Output price$50
AnthropicTextVision
claude-haiku-4-5
claude-haiku-4-5 is Anthropic's model at $1.00 input and $5.00 output per million tokens, with a 200K context window. On a typical 3:1 input-to-output mix that blends to roughly $2.00 per million tokens, placing it in the mid-range band of the 89 models tracked here — cheaper than 39% of them.
Context window200K
Input price$1 / 1M
Output price$5 / 1M
Output speed180 tok/s
Quality score82
Data updatedUpdated Jun 2026
Price verified2026-08-19 · Official
By price, claude-haiku-4-5 ranks 51st cheapest of 89 on input and 57th on output. Its quality score of 82 ranks 56th of 89, and on value (quality score ÷ output price) it ranks 52nd of 89. Its blended rate sits 35% below the catalog average.
In real billing terms: a support chatbot handling 50,000 turns a month (2,000 input / 500 output tokens each) costs about $225 per month on claude-haiku-4-5, while a RAG workload of 20,000 queries (8,000 input / 800 output tokens each) runs about $240. Because output is priced 5.0× input here, long-generation tasks scale cost fastest — cap output length before you optimise anything else.
A 200K context window holds roughly 267 pages of A4 text (at ~750 tokens per page), ranking 65th of 89 here. That covers typical RAG and multi-turn chat, though book-length inputs still need retrieval and chunking. Output runs at about 180 tok/s (15th fastest of 89), fast enough for real-time interaction and high-throughput batches.
We track 12 Anthropic models, and claude-haiku-4-5's blended rate sits 79% below that provider's average. If your task tolerates a little less headroom, qwen3.7-flash (Alibaba) blends about 97% cheaper at a quality score of 82, and is worth benchmarking first. Going the other way, Hunyuan HY3 scores 87 on quality for about 95% less.
The closest model in Anthropic's line-up is claude-haiku-5.5 ($0.10 input / $0.50 output, quality —). Against it, claude-haiku-4-5 scores 82 points higher on quality, has a 800K smaller context window, costs 90% more blended. When two models sit this close on price, the deciding factor is usually the quality ceiling — run both against your real prompts before picking.
Our read on it
180 tokens/second is its strongest number. Its price no longer matches the company it keeps at that speed.
Speed is the product
claude-haiku-4-5 streams at 180 tokens/second, well into the fast end of this catalog and level with deepseek-v4-flash and InclusionAI's Ling-3.0-flash — though raw throughput belongs to llama-3.3-70b-versatile on Groq at 300. Quality 82, 200K context, text and vision. What it is built for is unambiguous: interactive work — autocomplete, inline suggestions, streaming support replies, classification and routing.
In those jobs, time to first token matters far more than the quality of the five-hundredth. Users experience latency; they do not experience benchmark scores.
The pricing reality inside that speed band
The trouble is what else runs at 180 tokens/second. Ling-3.0-flash is $0.07/$0.22 with quality 80 and a 262K context. Against it, haiku-4-5 at $1.00/$5.00 costs roughly 14x more on input and 23x more on output, for two points of quality and a context window a quarter smaller. deepseek-v4-flash is $0.14/$0.28 at the same speed with quality 88 — again far cheaper and scoring higher.
The in-house comparison is not much kinder. claude-sonnet-5 is $2/$10 at quality 92 and 95 tokens/second. Haiku costs half as much and gives up ten points. That leaves it in a narrow gap: not cheap enough to be the workhorse for bulk processing, not strong enough to be the primary reasoning model.
Three reasons it still adds up
You are already on Anthropic. One key, one SDK, one tool-use format, one set of safety behaviours — mixing haiku with Sonnet 5 costs nothing in integration surface, and avoided complexity is a real saving even when it does not appear on an invoice.
Procurement and data residency. Enterprise buyers frequently have a completed vendor review for one provider, and the administrative cost of adding a second can exceed the difference on the bill. And 200K of context is genuinely ample for interactive work — most autocomplete and support flows never approach a million tokens. If none of those three apply to you, price deepseek-v4-flash and Ling-3.0-flash first, then decide whether the premium is buying anything.
Estimated monthly cost
Monthly API spend across four common workloads, next to the cheapest model we track.
| Workload | Usage assumption | This model | Cheapest model |
|---|---|---|---|
| Support chatbot | 2,000 input / 500 output tokens per turn, 50,000 turns per month | $225 | $3.00 |
| RAG document Q&A | 8,000 input / 800 output tokens per query, 20,000 queries per month | $240 | $3.84 |
| Coding assistant | 12,000 input / 2,000 output tokens per request, 10,000 requests per month | $220 | $3.20 |
| Batch summarization | 4,000 input / 400 output tokens per item, 200,000 items per month | $1,200 | $19 |
Derived directly from the rates on this page. Excludes prompt caching, batch discounts and free tiers.
How it compares
The alternatives worth benchmarking alongside it, from the same dataset.
| Model | Provider | Blended /1M | Quality | Price delta |
|---|---|---|---|---|
| claude-haiku-4-5 | Anthropic | $2.00 | 82 | — |
| qwen3.7-flashCheaper option | Alibaba | $0.055 | 82 | -97% |
| Hunyuan HY3Higher quality | Tencent | $0.10 | 87 | -95% |
| claude-haiku-5.5Same provider | Anthropic | $0.20 | — | -90% |
Best for
ChatbotAffordableFast
Modalities
TextVision
Where it fits — and where it doesn't
Good fit
- Chat and support: natural multi-turn conversation for user-facing products.
- High throughput and real-time UX: 180 tok/s keeps streamed responses feeling instant.
- Multimodal input: handles images, screenshots and scanned documents directly.
Poor fit
- No obvious disqualifiers — still benchmark on your own data before committing.
Frequently asked questions
How much does the claude-haiku-4-5 API cost?
claude-haiku-4-5 costs $1.00 per million input tokens and $5.00 per million output tokens. On a 3:1 input-to-output mix that blends to about $2.00 per million tokens.
What does claude-haiku-4-5 cost per month?
It depends on volume. Using the four worked scenarios on this page: a support chatbot runs about $225/month, RAG document Q&A about $240/month, a coding assistant about $220/month, and batch summarization about $1,200/month. Plug your own token counts into the cost calculator for a tailored figure.
How large is the claude-haiku-4-5 context window?
200K tokens — roughly 267 pages of A4 text at ~750 tokens per page, ranking 65th of the 89 models tracked here.
claude-haiku-4-5 or qwen3.7-flash — which is better value?
qwen3.7-flash blends to about $0.055 per million tokens versus $2.00 for claude-haiku-4-5 — roughly 97% cheaper — at a quality score of 82 against 82. If that quality gap does not show up on your task, take the cheaper one; if it does, stay with claude-haiku-4-5. Benchmark both on your own data before deciding.
What is claude-haiku-4-5 best suited for?
Tagged use cases: chatbot, cheap, fast. Chat and support: natural multi-turn conversation for user-facing products. Less suitable: No obvious disqualifiers — still benchmark on your own data before committing.
How current is this claude-haiku-4-5 pricing?
Pricing here was last verified on 2026-08-19 against official documentation. llmprice.app runs a daily collection job, but providers can change rates between runs — confirm against Anthropic's official pricing page before you commit.
Estimate the real cost of this model
Bring tokens per request and monthly volume into the calculator to compare total model costs.
Related models
Continue by provider and use case.
View model
claude-sonnet-4-6
Input price$3
Output price$15
claude-haiku-5.5
Input price$0.1
Output price$0.5
Pricing, speed and quality scores change over time. Confirm official documentation before production use.