Honest about its own work
Anthropic says it flags uncertainty more often, makes fewer unsupported claims and is about four times less likely than Opus 4.7 to let flaws in its own code slip by.
Anthropic
Operational· checked every few minutesClaude Opus 4.8 API
The Claude Opus 4.8 API at 52% below Anthropic's list price, on one key.
Price per 1M tokens52%off
- Input
- $2.40
$5.00 - Output
- $12.00
$25.00 - Cached input
- $0.24
$0.50
Model ID
claude-opus-4-8Updated chat.py
from openai import OpenAIclient = OpenAI( base_url="https://api.zurelay.com/v1", api_key="YOUR_ZURELAY_KEY",)stream = client.chat.completions.create( model="claude-opus-4-8", messages=[ {"role": "user", "content": "Explain quicksort briefly."}, ], stream=True,)for chunk in stream: if chunk.choices: print(chunk.choices[0].delta.content or "", end="")Base URL https://api.zurelay.com/v1. Works with any OpenAI SDK, and the Anthropic SDK and Claude Code. Chat API docs
- Context window
- 1M tokens
- Max output
- 128K tokens
- Reasoning
- low to max
- Input
- Text, Image
- Released
- May 28, 2026
- Status
- Operational
Pricing
Claude Opus 4.8 API pricing. A fraction of the list price.
Pay as you go from prepaid credit, with no subscription. Top up from $10. Requests that fail are never billed.
| Rate | Zurelay | Anthropic list | You save |
|---|---|---|---|
Input per 1M tokens | $2.40 | 52%off | |
Output per 1M tokens | $12.00 | 52%off | |
Cached input per 1M tokens | $0.24 | 52%off | |
Cache writes (5 min) per 1M tokens | $3.00 | 52%off | |
Cache writes (1 h) per 1M tokens | $4.80 | 52%off |
One rate at every prompt length, up to the full 1M context window. Streaming, tool calls and structured outputs cost nothing extra.
Every Claude model and priceSavings calculator
Move the sliders to your monthly usage.
Input tokens per month50M
Output tokens per month10M
Zurelay
$240.00
Anthropic list price
$500.00
You save every month$260.00
$3,120 a year
Cost
Claude Opus 4.8 API cost, worked out. And compared.
Typical workloads at Zurelay's live prices next to the list price, and what other places charge for the same model.
Chatbot
50,000 requests a month, 1,500 tokens in and 300 out, 30% of the input cached. $0.00623 a request.
$311.40a month$648.75at listDocuments
20,000 requests a month, 8,000 tokens in and 500 out, 20% of the input cached. $0.0217 a request.
$434.88a month$906.00at listCoding agent
5,000 requests a month, 40,000 tokens in and 1,500 out, 85% of the input cached. $0.0406 a request.
$202.80a month$422.50at listClassification
1,000,000 requests a month, 400 tokens in and 5 out. $0.00102 a request.
$1,020a month$2,125at list
Where to buy Claude Opus 4.8
Prices per 1M tokens
| Seller | Input | Output | vs Anthropic |
|---|---|---|---|
Anthropic List price, direct | $5.00 | $25.00 | — |
List price, plus 5.5% on card top-ups | $5.00 | $25.00 | +5.5% |
Zurelay No fee to add credit | $2.40 | $12.00 | 52%off |
OpenRouter's fee as its pricing page listed it on October 28, 2026. Zurelay's prices are live. Some labs also discount batch or off-peak requests.
Overview
What is Claude Opus 4.8?
Claude Opus 4.8 is the final Opus 4 model, released by Anthropic in May 2026 for agentic coding and knowledge work. The Zurelay Claude Opus 4.8 API serves the same model through OpenAI-compatible and Anthropic-compatible endpoints at 52% below Anthropic's list price. Teams keep it for its pinned behavior, its Opus 4.7 request shape and thinking that stays off until you ask for it.
Claude Opus 4.8 was released by Anthropic on May 28, 2026, six weeks after Opus 4.7, at the same list price. Anthropic lists its particular strengths as long-horizon agentic work, knowledge work, vision and memory tasks. On Zurelay, the Claude Opus 4.8 API uses the model ID claude-opus-4-8 and costs $2.40 per 1M input tokens and $12.00 per 1M output tokens, compared with Anthropic's list price of $5.00 and $25.00.
The headline change from Opus 4.7 is reliability. Anthropic says Opus 4.8 is around four times less likely to let flaws in code it wrote pass without comment, more likely to flag uncertainty about its own work, and less likely to make unsupported claims. Its prompting guide adds that it finds bugs with higher recall and precision than prior models in internal evals.
The specs: a 1M-token context window, up to 128K output tokens per request, text and image input, and text output. Reliable knowledge runs through January 2026. Thinking is adaptive and off unless you set it, and effort runs from low to max, including xhigh, with high as the default. Anthropic suggests starting at xhigh for coding and agentic work. The minimum cacheable prompt is 1,024 tokens. On Anthropic's Messages API, Opus 4.8 also added mid-conversation system messages, which update an agent's instructions without breaking the prompt cache.
Why pick Opus 4.8 over Opus 5? It takes exactly the same requests as Opus 4.7, so it is an easy step for 4.7 users. Without a thinking setting it answers directly, while Opus 5 thinks by default. It also keeps answers shorter and spawns fewer subagents than Opus 5, which helps when cost per request matters more than peak capability. Anthropic's current retirement date for Opus 4.8 is not sooner than May 28, 2027.
Strengths
Where Claude Opus 4.8 shines. And what teams build with it.
Better bug finding
Higher recall and precision on code review than earlier models in Anthropic's evals. Ask it to report every finding and filter in a later step.
Long autonomous runs
Built for long agentic work such as complex refactors. Give it the full task spec up front and run it at high or xhigh effort.
Drop-in for Opus 4.7
Same request surface as Opus 4.7, so upgrading is a model ID change plus light prompt tuning.
Use cases
Autonomous coding and refactors
Run long multi-step coding agents that plan, call tools and finish without constant correction.
Code review and QA
Review pull requests and generated code, with the model flagging what it is unsure about instead of glossing over it.
Knowledge work
Draft reports, analyses and structured extractions where a clear 'I am not sure' beats a confident guess.
Screenshots and documents
Read high-resolution screenshots, charts and scanned pages. For screen-driving agents, Anthropic suggests 1080p screenshots as a good balance of performance and cost.
Get started
Call Claude Opus 4.8 in three steps.
No waitlist and no new SDK. If your code already talks to OpenAI, it already talks to Zurelay.
- 1
Create a key
Sign up, add credit and create an API key. Set a monthly budget or a rate limit per key if you like.
- 2
Point your SDK at Zurelay
Change the base URL. Everything else in your code stays the same.
https://api.zurelay.com/v1 - 3
Use claude-opus-4-8
Send chat completions as usual. Streaming, tool calls and usage reporting work as you expect.
claude-opus-4-8
Works with the tools you already use
Compare
Similar models. All on the same key.
65%off
Claude Opus 5
$1.75 / $8.75 per 1M tokens
Pick Claude Opus 5 at the same list price for stronger results on hard coding, research and document work, if you accept longer answers and thinking on by default.
Claude Opus 5 API64%off
Claude Opus 4.7
$1.78 / $8.93 per 1M tokens
Stay on Claude Opus 4.7 only if your evals are pinned to its exact behavior; it takes the same requests but adds a larger tool-use system prompt to every request with tools.
Claude Opus 4.7 API65%off
Claude Opus 5.5
$1.40 / $7.00 per 1M tokens
Pick Claude Opus 5.5, Anthropic's current Opus at a lower list price, for new projects that can run with thinking always on.
Claude Opus 5.5 APIHow much does the Claude Opus 4.8 API cost on Zurelay?
Zurelay charges $2.40 per 1M input tokens and $12.00 per 1M output tokens, with cached input at $0.24. Anthropic's list price for Claude Opus 4.8 is $5.00 input and $25.00 output, so you save 52%. The rate is the same at every prompt length, and you pay from prepaid credit with no subscription.
Is there a cheaper Claude Opus 4.8 API than Anthropic's?
Yes. Zurelay serves claude-opus-4-8 at 52% below Anthropic's list price. It is the same model, not a smaller substitute. You keep your code and change the base URL and API key.
How do I call the Claude Opus 4.8 API with the OpenAI SDK?
Create a Zurelay API key, set the base URL to https://api.zurelay.com/v1 and set the model to claude-opus-4-8. Chat completions, streaming and tool calls use the request shapes you already know. Temperature, top_p and top_k have no effect: Opus 4.8 refuses them, so Zurelay leaves them out of the request.
Can I use Claude Opus 4.8 with the Anthropic SDK or Claude Code?
Yes. Set the Anthropic SDK base URL, or ANTHROPIC_BASE_URL for Claude Code, to https://api.zurelay.com with no /v1, because the client adds /v1/messages itself. Authenticate with your Zurelay key and select claude-opus-4-8. If you want thinking, send the adaptive thinking type; Opus 4.8 rejects the older budget_tokens form.
What is the Claude Opus 4.8 context window?
Claude Opus 4.8 has a 1M-token context window and returns up to 128K output tokens per request. On its tokenizer, 1M tokens is roughly 555k English words. At xhigh or max effort, Anthropic suggests a large output limit, starting around 64K tokens, and streaming the response.
What is Claude Opus 4.8 best at?
Long-horizon agentic work, knowledge work, vision and memory tasks, per Anthropic. It stands out on code review and on work where you want the model to say when it is unsure. It performs best when you give it the full task up front and run it at high or xhigh effort.
Should I use Claude Opus 4.8 or Claude Opus 5?
Both have the same list price. Opus 5 is stronger on hard coding, research and document work, but thinks by default, writes longer answers and delegates to subagents more. Opus 4.8 answers without thinking unless asked and matches the Opus 4.7 request shape, so it fits routes where you want shorter, cheaper replies or pinned behavior.
Does the Zurelay Claude Opus 4.8 API support streaming, and what are the limits?
Yes. Streaming is supported, and on chat completions you can set stream_options.include_usage to get token counts in the final chunk. Each key can carry its own monthly budget and requests-per-minute cap. If a route fails, Zurelay retries the request on another route, and failed requests are never billed.
Is Claude Opus 4.8 on Zurelay the same model as on Anthropic's API?
Yes. Requests to claude-opus-4-8 run on Claude Opus 4.8 itself, not a smaller or different model. Outputs are sampled, so exact wording can differ between calls on any API, including Anthropic's. Zurelay is an independent service and is not affiliated with or endorsed by Anthropic.
More models on Zurelay
All models- Zurelay Auto APIfrom $0.025 / $0.12 per 1M tokensAuto
- Claude Opus 5.5 API$1.40 / $7.00 per 1M tokens−65%
- GPT-6 Astra API$2.50 / $12.50 per 1M tokens−75%
- Nano Banana 2.1 APIfrom $0.03 per image−11%
- Nano Banana Pro APIfrom $0.035 per image−74%
- Nano Banana 2 APIfrom $0.025 per image−63%
- Nano Banana 2 Lite API$0.016 per image−52%
- Claude Sonnet 5.5 API$0.79 / $3.95 per 1M tokens−61%
- Claude Haiku 5.5 API$0.04 / $0.20 per 1M tokens−60%
- GPT-6.1 Sol API$0.50 / $2.50 per 1M tokens−75%
- GPT-6.1 Terra API$0.25 / $1.25 per 1M tokens−75%
- GPT-6 Sol API$0.50 / $2.50 per 1M tokens−75%
- GPT-6 Luna API$0.025 / $0.125 per 1M tokens−75%
- Grok 4.7 API$0.70 / $2.10 per 1M tokens−65%
- MiMo V2.6 Pro API$0.32 / $0.64 per 1M tokens−26%
- MiMo V2.6 Flash API$0.075 / $0.15 per 1M tokens−46%
- DeepSeek V4.1 Flash 1017 API$0.03 / $0.12 per 1M tokens−90%
- DeepSeek V4.1 Flash API$0.03 / $0.12 per 1M tokens−90%
- DeepSeek V4 Flash API$0.03 / $0.12 per 1M tokens−90%
- DeepSeek V4 Flash 0731 API$0.029 / $0.087 per 1M tokens−69%
- DeepSeek V4 Pro 0813 API$0.23 / $0.62 per 1M tokens−84%
- Claude Fable 5.2 API$3.20 / $16.00 per 1M tokens−60%
- Claude Fable 5.1 API$3.50 / $17.50 per 1M tokens−65%
- Claude Fable 5 API$3.50 / $17.50 per 1M tokens−65%
- Claude Opus 5 API$1.75 / $8.75 per 1M tokens−65%
- Claude Opus 4.7 API$1.78 / $8.93 per 1M tokens−64%
- Claude Opus 4.6 API$2.40 / $12.00 per 1M tokens−52%
- Claude Haiku 4.5 API$0.54 / $2.70 per 1M tokens−46%
- GPT-5.6 Sol API$1.14 / $5.70 per 1M tokens−72%
- GPT-5.6 Terra API$0.54 / $3.26 per 1M tokens−73%
- GPT-5.6 Luna API$0.08 / $0.49 per 1M tokens−59%
- Gemini 4 Hydrogen API$0.60 / $3.00 per 1M tokens−60%
- Gemini 4 Argon API$0.80 / $4.00 per 1M tokens−60%
- Gemini 3.1 Pro API$0.50 / $3.45 per 1M tokens−71%
- Gemini 3.8 Flash API$0.22 / $1.10 per 1M tokens−71%
- Gemini 3.7 Flash API$0.32 / $1.59 per 1M tokens−58%
- Gemini 3.5 Flash-Lite API$0.12 / $1.04 per 1M tokens−58%
- Gemini 3.1 Flash-Lite API$0.10 / $0.59 per 1M tokens−61%
- GLM-5.3 API$0.38 / $1.21 per 1M tokens−73%
- GLM-5.2 API$0.38 / $1.21 per 1M tokens−73%
- GLM-5.3 Flash API$0.058 / $0.19 per 1M tokens−62%
- Kimi K3.1 Pro API$0.80 / $4.00 per 1M tokens−60%
- Kimi K3.1 Pro Highspeed API$3.30 / $16.50 per 1M tokens−18%
- Kimi K3 API$1.27 / $6.37 per 1M tokens−58%
- Qwen 3.8 Max API$1.20 / $3.60 per 1M tokens−40%
- Qwen 3.8 Flash API$0.07 / $0.22 per 1M tokens−53%
- MiniMax M3.1 Flash API$0.06 / $0.24 per 1M tokens−60%
- Qwen 3.8 Omni Flash API$0.075 / $0.235 per 1M tokens−50%
- HY4 Preview API$0.40 / $1.20 per 1M tokens−52%
- MiniMax M3 API$0.13 / $0.51 per 1M tokens−57%
- Step 5 Preview API$0.25 / $0.68 per 1M tokens−75%
- GPT Image 2.6 Flare APIfrom $0.008 per image−85%
- GPT Image 2.6 Sunburst APIfrom $0.008 per image−85%
- GPT Image 2.5 Flare APIfrom $0.008 per image−85%
- GPT Image 2.5 Sunburst APIfrom $0.008 per image−85%
- GPT Image 2 API$0.02 per image−62%
- Grok Imagine Image 2.0 APIfrom $0.045 per image−25%
- FLUX 3 Image APIfrom $0.04 per image−17%
- Seedance 2.5 APIfrom $0.135 per secondVideo
- Seedance 2.0 APIfrom $0.097 per secondVideo
- Seedance 2.0 Fast APIfrom $0.074 per secondVideo
- Kling 4.0 APIfrom $0.15 per secondVideo
- Seedance 2.0 Mini APIfrom $0.047 per secondVideo
- Gemini Omni 1.2 Flash APIfrom $0.042 per secondVideo
- Kling 3.0 APIfrom $0.125 per secondVideo
- Gemini Omni Flash APIfrom $0.042 per secondVideo
- Wan 3.0 Prime APIfrom $0.06 per secondVideo
- Wan 3.0 APIfrom $0.039 per secondVideo
- Grok Imagine Video 1.5 APIfrom $0.071 per secondVideo
- Veo 3.1 Fast APIfrom $0.025 per secondVideo
- Muse Video APIfrom $0.02 per secondVideo
- Vidu Q3 Turbo APIfrom $0.03 per secondVideo
- PixVerse V6 APIfrom $0.019 per secondVideo
- MiniMax H3 APIfrom $0.039 per secondVideo
- Suno V6 API$0.096 per song−0%
- Gemini 3.8 Flash TTS API$0.30 / $5.40 per 1M tokens−40%
- Gemini 3.1 Flash TTS API$0.60 / $12.00 per 1M tokens−40%
- GPT-6 Luna Decisions API$0.075 per 1M input tokens−25%
- Jev 1.14 API$0.035 per 1M input tokens−17%
- Jev 1.13 API$0.035 per 1M input tokens−17%
- Kimi K2.6 API$0.30 / $1.30 per 1M tokens−68%
- Kimi K2.5 API$0.20 / $0.95 per 1M tokens−68%
- MiniMax M2.5 API$0.10 / $0.39 per 1M tokens−68%
- Ministral 3 3B API$0.03 / $0.03 per 1M tokens−70%
- Ministral 3 14B API$0.06 / $0.06 per 1M tokens−70%
- Abliterated Large V2 API$2.93 / $4.88 per 1M tokens−2%
- Text Embedding 3 Small API$0.01 per 1M input tokens−50%
- Text Embedding 3 Large API$0.065 per 1M input tokens−50%
- GLM-4.7 Flash Heretic API$0.061 / $0.35 per 1M tokens−13%
- Gemma 4 Uncensored API$0.155 / $0.48 per 1M tokens−4%
Stop paying list price.
Start saving today.
Create a key in seconds, point the SDK you already use at Zurelay, and every request costs up to 90% less from the first token.