claude-fable-5.1
Input price$10
Output price$50
AnthropicTextVision
claude-opus-5
claude-opus-5 is Anthropic's model at $5.00 input and $25.00 output per million tokens, with a 1049K context window. On a typical 3:1 input-to-output mix that blends to roughly $10.00 per million tokens, placing it in the premium band of the 76 models tracked here — cheaper than 7% of them.
Context window1049K
Input price$5 / 1M
Output price$25 / 1M
Output speed58 tok/s
Quality score97
Data updatedUpdated Aug 2026
Price verified2026-08-19 · Official
By price, claude-opus-5 ranks 69th cheapest of 76 on input and 70th on output. Its quality score of 97 ranks 2nd of 76, and on value (quality score ÷ output price) it ranks 60th of 76. Its blended rate sits 233% above 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 $1,125 per month on claude-opus-5, while a RAG workload of 20,000 queries (8,000 input / 800 output tokens each) runs about $1,200. Because output is priced 5.0× input here, long-generation tasks scale cost fastest — cap output length before you optimise anything else.
A 1049K context window holds roughly 1,399 pages of A4 text (at ~750 tokens per page), ranking 3rd of 76 here. That is enough to drop in an entire technical manual or several source files at once, which suits long-document summarization and codebase-wide analysis. Output runs at about 58 tok/s (59th fastest of 76), which is better suited to asynchronous or background jobs.
We track 9 Anthropic models, and claude-opus-5's blended rate sits 12% below that provider's average. If your task tolerates a little less headroom, qwen3.8-max (Alibaba) blends about 70% cheaper at a quality score of 94, and is worth benchmarking first. Going the other way, gpt-5.6-sol scores 98 on quality for about 20% less.
The closest model in Anthropic's line-up is claude-opus-4-8 ($5.00 input / $25.00 output, quality 94). Against it, claude-opus-5 scores 3 points higher on quality, runs 8 tok/s faster (about 16%). 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
Anthropic cut flagship pricing to $5/$25, and that changes when a flagship is the right tool — not just whether you can afford one.
The moment flagship stopped meaning luxury
Our own price history has claude-opus-4-8 at $15/$75 in January 2026, $10/$50 in April, and $5/$25 by June. claude-opus-5 entered the catalog in August at that same $5/$25, scoring 97. Today's flagship rate is a third of what it was at the start of the year, and the capability moved the other way.
The strategic intent is not subtle. It moves running a flagship on production traffic from a decision that needs finance sign-off to one an engineer can make. $25 per 1M output tokens is still real money, but it now sits in the band where you can put the model in an agent's main loop rather than reserving it for the last mile.
What it means that coding, reasoning and long-context all apply
Opus 5 carries all three use-case tags at once, which is uncommon in this catalog — most models claim one or two. Combined with 1,049K of context, that points at a specific shape of work: read an entire repository, a full specification, a whole issue thread, then make a change that requires reasoning across all of it.
The price is 58 tokens/second, toward the slow end of what we track (claude-haiku-4-5 does 180). That is a trade, not a defect. Tasks that need a global understanding built across a million tokens were never going to be won on time-to-first-token. Anything with a user waiting on a response should not have Opus 5 in front of it.
When to skip Sonnet 5 and go straight to Opus 5
Its sibling claude-sonnet-5 is $2/$10, quality 92, the same 1,049K context, and faster at 95 tokens/second. Five points of quality for 2.5x the price: for most applications Sonnet 5 is still the rational answer, and it got harder to argue with once Anthropic cancelled the September increase and left Sonnet at $2/$10.
Two cases flip it. First, deeply autonomous agents, where an error at step three compounds through every step after it and the cost of extra correction rounds dwarfs the per-token difference. Second, output that lands in an irreversible process — a PR opened automatically, a document sent to a client, a structured extraction written straight to a database. What those share is the absence of a human checkpoint. Where a review gate does exist, Sonnet 5 plus the saved budget spent on extra self-checking passes usually produces better results than a single Opus 5 call.
One more thing has to be said plainly. On 22 August OpenAI moved gpt-5.6-sol to $4/$20 at quality 98. It now beats Opus 5 on input rate, output rate and score, matches its 1,049K context and adds audio. What remains in Opus 5's favour is everything a price table cannot show — tool-use reliability, the sunk cost of prompts already tuned against it, and not routing every flagship request through one vendor. Those are real arguments. They just have to be made out loud now, because the pricing one is gone.
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 | $1,125 | $3.00 |
| RAG document Q&A | 8,000 input / 800 output tokens per query, 20,000 queries per month | $1,200 | $3.84 |
| Coding assistant | 12,000 input / 2,000 output tokens per request, 10,000 requests per month | $1,100 | $3.20 |
| Batch summarization | 4,000 input / 400 output tokens per item, 200,000 items per month | $6,000 | $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-opus-5 | Anthropic | $10.00 | 97 | — |
| qwen3.8-maxCheaper option | Alibaba | $3.00 | 94 | -70% |
| gpt-5.6-solHigher quality | OpenAI | $8.00 | 98 | -20% |
| claude-opus-4-8Same provider | Anthropic | $10.00 | 94 | +0% |
| qwen3.5Best value overall | Alibaba | $0.11 | 91 | -99% |
Best for
CodingReasoningLong Context
Modalities
TextVision
Where it fits — and where it doesn't
Good fit
- Coding: writing, refactoring and debugging across multi-file project context.
- Reasoning: multi-step decomposition, planning and tasks that need rigorous inference.
- Long documents: a 1049K window ingests a full contract, paper or large codebase in one pass.
- Multimodal input: handles images, screenshots and scanned documents directly.
Poor fit
- High-frequency batch jobs: at $10.00 blended per million tokens, bills scale quickly at volume.
- Latency-critical interactive use: 58 tok/s makes long responses feel sluggish.
Frequently asked questions
How much does the claude-opus-5 API cost?
claude-opus-5 costs $5.00 per million input tokens and $25.00 per million output tokens. On a 3:1 input-to-output mix that blends to about $10.00 per million tokens.
What does claude-opus-5 cost per month?
It depends on volume. Using the four worked scenarios on this page: a support chatbot runs about $1,125/month, RAG document Q&A about $1,200/month, a coding assistant about $1,100/month, and batch summarization about $6,000/month. Plug your own token counts into the cost calculator for a tailored figure.
How large is the claude-opus-5 context window?
1049K tokens — roughly 1,399 pages of A4 text at ~750 tokens per page, ranking 3rd of the 76 models tracked here.
claude-opus-5 or qwen3.8-max — which is better value?
qwen3.8-max blends to about $3.00 per million tokens versus $10.00 for claude-opus-5 — roughly 70% cheaper — at a quality score of 94 against 97. If that quality gap does not show up on your task, take the cheaper one; if it does, stay with claude-opus-5. Benchmark both on your own data before deciding.
What is claude-opus-5 best suited for?
Tagged use cases: coding, reasoning, long-context. Coding: writing, refactoring and debugging across multi-file project context. Less suitable: High-frequency batch jobs: at $10.00 blended per million tokens, bills scale quickly at volume.
How current is this claude-opus-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-opus-4-8
Input price$5
Output price$25
claude-opus-4-6
Input price$5
Output price$25
Pricing, speed and quality scores change over time. Confirm official documentation before production use.