Claude Opus 5 vs DeepSeek V4 Pro：價格與規格實測

什麼情況選哪一個

兩者共享 1000000 token 上下文且都能關閉思考，所以差別在模態和價目表：claude-opus-5 在文字之外接受圖像輸入並帶明確的思考能力，而 deepseek-v4-pro 只處理文字，但輸入便宜約 3.8 倍（$1.32 對 $5）、輸出便宜約 6.3 倍（$3.96 對 $25）、快取讀取便宜約 3.8 倍（$0.132 對 $0.5）。當提示詞裡包含截圖或圖表時選 claude-opus-5；大批量文字和程式碼工作選 deepseek-v4-pro，它 393216 token 的最大輸出也超過 128000 的上限。

Benchmark 成績

高於平均沒有模型更高Claude Opus 538 ~~/ 46~~8 ~~/ 46~~DeepSeek V4 Pro5 ~~/ 10~~0 ~~/ 10~~

Claude Opus 5 DeepSeek V4 Pro 其他有成績的模型其他模型平均 ★ 沒有模型分數更高

NL2Repo

沒有模型分數更高 75.3%

35.5%

BioMysteryBench hard

49.4%

N/A

OSWorld 2.0

沒有模型分數更高 70.6%

N/A

ExploitBench (Cap%)

70%

N/A

HealthBench Professional

59.8%

N/A

Finance Agent v2

58.6%

N/A

Legal Agent Benchmark

6.7%

N/A

Humanity's Last Exam no tools

56.6%

37.7%

BrowseComp

90.8%

N/A

LVBench

75.4%

N/A

Humanity's Last Exam no tools

56.6%

37.7%

GPQA Diamond

93.4%

N/A

Humanity's Last Exam with tools

63.6%

N/A

HLE-Verified

54.4%

N/A

FrontierMath Tier 4 (v2)

73.2%

N/A

ARC-AGI-3

30.2%

N/A

ARC-AGI-1

97.5%

N/A

ARC-AGI-2

90.4%

N/A

GPQA Diamond max

N/A

90.1%

Apex Shortlist high

N/A

85.5%

Apex Shortlist max

N/A

90.2%

Chinese-SimpleQA high

N/A

77.7%

Chinese-SimpleQA max

N/A

84.4%

CorpusQA 1M high

N/A

56.5%

CorpusQA 1M max

N/A

62%

GPQA Diamond high

N/A

89.1%

HMMT Feb. 2026 high

N/A

94%

HMMT Feb. 2026 max

N/A

95.2%

Humanity's Last Exam no tools, high

N/A

34.5%

Humanity's Last Exam no tools, max

N/A

37.7%

Humanity's Last Exam with tools, high

N/A

44.7%

Humanity's Last Exam with tools, max

N/A

48.2%

IFBench max

N/A

77%

IMOAnswerBench high

N/A

88%

IMOAnswerBench max

N/A

89.8%

MMLU-Pro high

N/A

87.1%

MMLU-Pro max

N/A

87.5%

MMMLU max

N/A

87.9%

MRCR 1M high

N/A

83.3%

MRCR 1M max

N/A

83.5%

SimpleQA-Verified high

N/A

46.2%

SimpleQA-Verified max

N/A

57.9%

Apex Shortlist non-think

N/A

9.2%

Chinese-SimpleQA non-think

N/A

75.8%

CorpusQA 1M non-think

N/A

35.6%

CritPt max

N/A

12.9%

Global PIQA max

N/A

90.5%

GPQA Diamond non-think

N/A

72.9%

HMMT Feb. 2026 non-think

N/A

31.7%

Humanity's Last Exam no tools, non-think

N/A

7.7%

IFEval max

N/A

91.9%

IMOAnswerBench non-think

N/A

35.3%

INCLUDE max

N/A

86.1%

MAXIFE max

N/A

88.9%

MMLU-Pro non-think

N/A

82.9%

MMLU-ProX max

N/A

83.9%

MMLU-Redux max

N/A

94.8%

MRCR 1M non-think

N/A

44.7%

MRCR v2 128K max

N/A

74.4%

NOVA-63 max

N/A

52.8%

PolyMATH max

N/A

72%

QwenWorldBench max

N/A

52.3%

SimpleQA-Verified non-think

N/A

45%

SuperGPQA max

N/A

69.9%

WMT24++ max

N/A

82.2%

NL2Repo

沒有模型分數更高 75.3%

35.5%

SWE-Bench Pro

N/A

59%

DeepSWE 1.1

68.8%

N/A

Terminal-Bench 2.1

89.1%

N/A

SWE-bench Multilingual

N/A

76.2%

Terminal-bench 4.0

52.3%

N/A

Terminal-bench 3.0

沒有模型分數更高 43.3%

N/A

FrontierCode 1.1 Main

48%

N/A

ProgramBench Almost Solved

沒有模型分數更高 37%

N/A

Terminal Bench 2.0 Terminus 2

N/A

67.9%

FrontierCode 1.1 Extended

63.6%

N/A

CursorBench 3.2.0

70%

N/A

CursorBench 4.0

46.6%

N/A

Frontier-Bench v0.1

沒有模型分數更高 43.3%

N/A

SWE-Bench Pro max

N/A

55.4%

Codeforces (Elo) high · 2816-3052 據 DeepSeek 公布 · 2026-08-26

N/A

2919

LiveCodeBench high

N/A

89.8%

LiveCodeBench max

N/A

93.5%

SWE-bench Multilingual max

N/A

76.2%

SWE-Bench Pro high

N/A

54.4%

SWE-bench Verified high

N/A

79.4%

SWE-bench Verified max

N/A

80.6%

Terminal Bench 2.0 high

N/A

63.3%

Terminal Bench 2.0 max

N/A

67.9%

Terminal Bench 2.0 Terminus 2, max

N/A

67.9%

Codeforces (Elo) max · 3052-3206 據 DeepSeek 公布 · 2026-08-26

N/A

3206

LiveCodeBench non-think

N/A

56.8%

NL2Repo max

N/A

35.5%

QwenSVGBench Elo, max · 1506-1541 據 Alibaba (Qwen) 公布 · 2026-08-26

N/A

1506

QwenWebDev (Elo) Elo, max · 1570-1617 據 Alibaba (Qwen) 公布 · 2026-08-26

N/A

1570

SWE-bench Multilingual non-think

N/A

69.8%

SWE-bench Multilingual high

N/A

74.1%

SWE-Bench Pro non-think

N/A

52.1%

SWE-bench Verified non-think

N/A

73.6%

Terminal Bench 2.0 non-think

N/A

59.1%

供應商公布： Alibaba (Qwen) Anthropic ByteDance DeepSeek Google MiniMax Moonshot OpenAI Tencent Z.ai

定價

	Claude Opus 5	DeepSeek V4 Pro	Δ
輸入 / 1M tokens	$5	$1.32	3.8×
輸出 / 1M tokens	$25	$3.96	6.3×
快取讀取 / 1M tokens	$0.5	$0.132	3.8×
快取寫入	1.25x (5m) / 2x (1h)	不額外計費	-

費率取自網站建置時的即時目錄；各模型頁面都列有最新的價目。

兩者的相對位置：每 1M tokens 的輸入價格，涵蓋同一計費單位下全部 76 個聊天模型（對數尺度）

Claude Opus 5 · $5 DeepSeek V4 Pro · $1.32

$0.05 · Qwen3 VL Flash $30 · GPT-5.4 Pro

功能

	Claude Opus 5	DeepSeek V4 Pro
工具使用	是	是
思考控制	可設定	可設定
結構化輸出	是	是
提示詞快取	顯式（由你標記前綴）	隱式（自動）
快取存活時間	5m default, 1h option	no fixed TTL (evicted when unused)
最小快取前綴	1024 個 token	未公布

規格

	Claude Opus 5	DeepSeek V4 Pro
輸入模態	文字圖像	文字
輸出模態	文字	文字
發布日期	2026-07-24	2026-04-24
知識截止日期	2026-05	-
上下文視窗	1M	1M
最大輸出	128K	393K
思考參數	thinking.type output_config.effort	thinking.type reasoning_effort
可接受的值	thinking.type adaptive disabled effort low medium high xhigh max	thinking.type enabled disabled reasoning_effort high max (low and medium map to high, xhigh maps to max)
預設值	thinking on effort high (Claude API and Claude Code)	enabled, with reasoning_effort high some complex agent requests are automatically set to max

規格摘錄自各供應商的文件；供應商沒有公布的項目就直接略過，不自行推測。完整來源： Claude Opus 5 · DeepSeek V4 Pro

同一段提示詞，兩個模型，經閘道實測

寫給 CFO 的推理修正有 bug 的函式依 schema 擷取資料依照限制條件撰寫

提示詞 Explain to a CFO, in exactly three sentences, why our LLM bill doubled after we enabled reasoning. 檢查恰好 3 句

Claude Opus 5 通過 · 3 sentences

When reasoning is enabled, the model generates a hidden chain of "thinking" tokens before it writes the visible answer, and those internal tokens are billed at the output rate even though no one ever sees them — commonly 3–10x more tokens than the response itself. Because output tokens are priced several times higher than input tokens, that added volume hits us at our most expensive rate, which is why the bill roughly doubled while our request count stayed flat. The spend is controllable: we can cap the reasoning budget per call, route only genuinely hard requests to reasoning mode and cheaper models for the rest, and track cost per successfully completed task rather than cost per token, since a single correct answer usually beats three cheap retries plus human cleanup.

輸出 529 tok 延遲 9.0 s

DeepSeek V4 Pro 通過 · 3 sentences

Enabling reasoning causes the model to generate a lengthy internal chain‑of‑thought before it delivers a visible answer, and we are billed for every one of those hidden reasoning tokens just as we are for the final output. In most interactions, the volume of these unseen tokens at least equals the visible response, effectively doubling the total tokens consumed per query. That doubled token consumption translates directly into a doubled per‑query cost, which is why our overall LLM bill doubled.

輸出 641 tok (+545 思考) 延遲 11.5 s

是否照指令做（剛好三句，數得出來）、語氣是否貼合對象（寫給 CFO 看），以及下方 token 計量顯示出的隱藏思考計費落差。

提示詞 This function is supposed to return the longest run of consecutive equal items, but callers report it is off by one on single-element inputs and misses runs that end at the last index. Fix it and explain the bug in one paragraph. 檢查修正後通過測試

Claude Opus 5 通過 · 8/8 cases

## Fixed version ```python def longest_run(xs): if not xs: return 0 best = 1 cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 else: cur = 1 if cur > best: best = cur return best ``` ## The bug The counters were tracking the wrong quantity: `cur` counted *matching adjacent pairs* rather than *elements in the current run*, and both counters started at `0` instead of `1`. Since a run of length *n* contains only *n − 1* adjacent equal pairs, every result came back one too small — `[1, 1]` returned 1, `[3, 3, 3]` returned 2 — and a single-element list returned 0 because the loop body never executed and `best` kept its initial value of `0`. […]

輸出 1094 tok 延遲 16.1 s

DeepSeek V4 Pro 通過 · 8/8 cases

```python def longest_run(xs): if not xs: return 0 best = 1 cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 best = max(best, cur) else: cur = 1 return best ``` **Explanation:** The bug stems from initializing `cur = 0` and only incrementing it when two adjacent elements match, which counts the number of *additional* equal elements rather than the actual run length (which is `cur + 1`). As a result, the returned `best` is always one less than the true longest run—most obviously returning 0 for a single-element list instead of 1. […]

輸出 2418 tok (+2131 思考) 延遲 35.6 s

修正是否真的正確（跑得起來）、說明的資訊密度，以及在範圍明確的任務上的 token 效率。

提示詞 Extract a JSON object with fields {company, quarter, revenue_usd_m, growth_pct, guidance} from this text. Use null for anything not stated; add no other fields. 檢查有效的 JSON，完全符合 schema

Claude Opus 5 通過 · 5/5 fields, guidance "Withheld pending Q4 close"

```json { "company": "Meridian Robotics", "quarter": "Q3", "revenue_usd_m": 412, "growth_pct": 11, "guidance": "Withheld pending Q4 close" } ``` Notes on judgment calls: - **quarter**: "September quarter" is labeled Q3 because the text refers to an upcoming "Q4 close," implying a calendar-aligned fiscal year. No fiscal year is given. - **revenue_usd_m**: 412 is stated as constant-currency and hedged ("call it"), so it's approximate; the currency is not explicitly stated as USD. - **growth_pct**: "up eleven and change" means slightly more than 11, so 11 is a floor rather than an exact figure.

輸出 539 tok 延遲 7.6 s

DeepSeek V4 Pro 通過 · 5/5 fields, guidance null

{ "company": "Meridian Robotics", "quarter": "September quarter", "revenue_usd_m": 412, "growth_pct": "eleven and change", "guidance": null }

輸出 1204 tok (+1153 思考) 延遲 19.9 s

是否遵守 schema（不自創欄位）、面對幻覺誘因的表現（題目明確不給 guidance），以及結構化輸出路徑的差異。

提示詞 Write a 120-word product announcement for an API gateway feature that caches prompts across providers. Forbidden words: "seamless", "unlock", "game-changer", "revolutionize", "empower". Exactly one sentence must be a question. 檢查 120 詞，0 個禁用詞

Claude Opus 5 通過 · 123 words, 0 banned, 1 question

**Cross-Provider Prompt Caching** We're introducing Cross-Provider Prompt Caching in the API Gateway. Repeated system prompts, long documents, and few-shot examples are stored once at the gateway layer and reused across OpenAI, Anthropic, Google, and self-hosted models. Instead of paying full input token costs on every request, your application sends a cache reference, and the gateway rehydrates the context before forwarding it downstream. Why does that matter? Teams running high-volume agents and retrieval pipelines typically see input token spend fall 40 to 70 percent, with median latency dropping by several hundred milliseconds. Caches are scoped per project, encrypted at rest, and invalidated automatically when a prompt template changes. Enable it with a single header, and see the docs for TTL tuning and per-route controls.

輸出 1593 tok 延遲 19.1 s

DeepSeek V4 Pro 沒有可評分的回答 · no answer text within 16,384 tokens (all of it went to thinking)

模型沒有回傳任何回答文字，token 額度全都用在隱藏的思考上了。

輸出 8193 tok (+8192 思考) 延遲 106.2 s

是否遵守限制（字數上限、禁用詞清單、只能有一句問句）、文字風格的特徵，以及長度控制。

改一行程式碼就能在兩者之間切換

下面每個頁籤都列了這兩個模型 ID，要改的只有醒目標示的那兩行。端點、金鑰和請求格式都不變。

from openai import OpenAI

client = OpenAI(
    base_url="https://synthorai.io/v1",
    api_key="sk-syn-...",
)

resp = client.chat.completions.create(
    model="claude-opus-5",
    # model="deepseek-v4-pro",  # 取消這一行的註解，並把上一行註解掉
    messages=[{"role": "user", "content": "Summarize this diff"}],
    reasoning_effort="medium",
)
print(resp.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://synthorai.io/v1",
  apiKey: "sk-syn-...",
});

const resp = await client.chat.completions.create({
  model: "claude-opus-5",
  // model: "deepseek-v4-pro",  // 取消這一行的註解，並把上一行註解掉
  messages: [{ role: "user", content: "Summarize this diff" }],
  reasoning_effort: "medium",
});
console.log(resp.choices[0].message.content);

curl https://synthorai.io/v1/chat/completions \
  -H "Authorization: Bearer sk-syn-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-opus-5",
    # "model": "deepseek-v4-pro",  # 取消這一行的註解，並把上一行註解掉
    "messages": [{"role": "user", "content": "Hello"}],
    "reasoning_effort": "medium"
  }'

package main

import (
	"context"
	"fmt"

	"github.com/openai/openai-go/v3"
	"github.com/openai/openai-go/v3/option"
)

func main() {
	client := openai.NewClient(
		option.WithBaseURL("https://synthorai.io/v1"),
		option.WithAPIKey("sk-syn-..."),
	)
	resp, _ := client.Chat.Completions.New(context.TODO(), openai.ChatCompletionNewParams{
		Model: "claude-opus-5",
		// Model: "deepseek-v4-pro",  // 取消這一行的註解，並把上一行註解掉
		Messages: []openai.ChatCompletionMessageParamUnion{
			openai.UserMessage("Summarize this diff"),
		},
		ReasoningEffort: openai.ReasoningEffortMedium,
	})
	fmt.Println(resp.Choices[0].Message.Content)
}

import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.chat.completions.*;
import com.openai.models.ReasoningEffort;

OpenAIClient client = OpenAIOkHttpClient.builder()
    .baseUrl("https://synthorai.io/v1")
    .apiKey("sk-syn-...")
    .build();

ChatCompletion resp = client.chat().completions().create(
    ChatCompletionCreateParams.builder()
        .model("claude-opus-5")
        // .model("deepseek-v4-pro")  // 取消這一行的註解，並把上一行註解掉
        .addUserMessage("Summarize this diff")
        .reasoningEffort(ReasoningEffort.MEDIUM)
        .build());
System.out.println(resp.choices().get(0).message().content().orElse(""));

取得 API 金鑰 →

常見問題

Claude Opus 5 和 DeepSeek V4 Pro 哪個比較便宜？

以「輸入 / 1M tokens」來看，DeepSeek V4 Pro 比較便宜（$1.32 對 $5，相差 3.8×）。其他項目的結果可能相反，完整價目請看上表；實際成本要看你的用量組合。

可以只串接一次，就對 Claude Opus 5 和 DeepSeek V4 Pro 做 A/B 測試嗎？

可以。兩個模型都走同一個 OpenAI 相容端點，用的也是同一把 API 金鑰，切換時只要改一行裡的模型名稱字串。你可以把一部分流量分別導到兩邊，再直接比較帳單。

Claude Opus 5 與 DeepSeek V4 Pro 支援提示詞快取嗎？

支援。兩者的快取讀取費率都低於輸入費率，所以前綴已經進快取的工作負載，實際成本會比官網價算出來的低。確切的快取讀取價格請見上方定價表。

Benchmark 成績

定價

功能

規格

同一段提示詞，兩個模型，經閘道實測

改一行程式碼就能在兩者之間切換

常見問題

相關比較