DeepSeek V4.1 Flash 對比 Gemini 3.7 Flash:價格與規格實測

Q: DeepSeek V4.1 Flash 和 Gemini 3.7 Flash 哪個比較便宜？

DeepSeek V4.1 Flash 在 輸入 / 1m tokens 上較便宜（$0.3 對比 $0.75，相差 2.5×）。其他項目可能呈現相反結果 — 上表提供完整資訊，實際成本取決於您的使用組合。

何時用哪一個

兩者皆為文字輸出模型，上下文長度大致相同（deepseek-v4.1-flash 為 1000000 token，gemini-3.7-flash 為 1048576），因此真正的差異在於價格、輸出長度與輸入模態。若要進行大批量的文字與圖片作業，請選擇 deepseek-v4.1-flash：其 $0.3 的輸入與 $1.2 的輸出，比 Gemini 的 $0.75 與 $3.75 分別便宜了 2.5x 與大約 3.1x，且其 393216-token 的最大輸出是 65536 上限的六倍，這對於長篇生成至關重要。當您需要音訊或影片輸入時，請選擇 gemini-3.7-flash，其定價為每百萬個音訊輸入 token $0.75，而快取讀取則為 $0.075。

Benchmark 成績

領先高於同儕均值無人分數更高DeepSeek V4.1 Flash516 ~~/ 19~~4 ~~/ 19~~Gemini 3.7 Flash017 ~~/ 24~~3 ~~/ 24~~

雙方都被測過的 5 項。

DeepSeek V4.1 Flash Gemini 3.7 Flash 其他被測模型同儕均值 ★ 無人分數更高

Terminal-Bench 2.1

無人分數更高 90.6%

85.8%

BioMysteryBench hard

N/A

43.5%

OSWorld 2.0

N/A

47.9%

Cybergym

無人分數更高 88.1%

N/A

Finance Agent v2

N/A

59%

Harvey Lab-AA

N/A

90.7%

GPQA Diamond

90.9%

N/A

AutomationBench (v1.0.6)

無人分數更高 54.8%

52.3%

LVBench

N/A

無人分數更高 85.4%

Terminal-Bench 2.1

無人分數更高 90.6%

85.8%

DeepSWE 1.1

無人分數更高 74.2%

65.3%

Terminal-bench 3.0

30%

14.9%

Terminal-bench 4.0

31.2%

11.2%

NL2Repo

64%

N/A

ProgramBench Almost Solved

20.3%

N/A

FrontierCode 1.1 Main

N/A

43.6%

Code Arena Elo · 1523-1588 據 Google 公布 · 2026-08-25

N/A

無人分數更高 1588

Codeforces (Elo) max · 3289-3471 據 DeepSeek 公布 · 2026-09-14

3471

N/A

DeepSWE 1.1 Claude Code 據 DeepSeek 公布

沒有其他模型的公開成績

69.8%

N/A

DeepSWE 1.1 Codex 據 DeepSeek 公布

沒有其他模型的公開成績

65.6%

N/A

DeepSWE 1.1 DSH Minimal 據 DeepSeek 公布

沒有其他模型的公開成績

72.6%

N/A

DeepSWE 1.1 DSH PTC 據 DeepSeek 公布

沒有其他模型的公開成績

67.6%

N/A

DeepSWE 1.1 DSH Standard 據 DeepSeek 公布

沒有其他模型的公開成績

70.5%

N/A

DeepSWE 1.1 mini-SWE 據 DeepSeek 公布

沒有其他模型的公開成績

74.2%

N/A

DeepSWE 1.1 OpenCode 據 DeepSeek 公布

沒有其他模型的公開成績

65.5%

N/A

DeepSWE 1.1 Pi 據 DeepSeek 公布

沒有其他模型的公開成績

66.2%

N/A

Terminal-Bench 2.1 Claude Code 據 DeepSeek 公布

沒有其他模型的公開成績

88%

N/A

Terminal-Bench 2.1 Codex 據 DeepSeek 公布

沒有其他模型的公開成績

84.1%

N/A

Terminal-Bench 2.1 DSH Minimal 據 DeepSeek 公布

沒有其他模型的公開成績

90.6%

N/A

Terminal-Bench 2.1 DSH PTC 據 DeepSeek 公布

沒有其他模型的公開成績

85.8%

N/A

Terminal-Bench 2.1 DSH Standard 據 DeepSeek 公布

沒有其他模型的公開成績

85.8%

N/A

Terminal-Bench 2.1 mini-SWE 據 DeepSeek 公布

沒有其他模型的公開成績

90.3%

N/A

Terminal-Bench 2.1 OpenCode 據 DeepSeek 公布

沒有其他模型的公開成績

85%

N/A

Terminal-Bench 2.1 Pi 據 DeepSeek 公布

沒有其他模型的公開成績

86.1%

N/A

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

定價

	DeepSeek V4.1 Flash	Gemini 3.7 Flash	Δ
輸入 / 1M tokens	$0.3	$0.75	0.4×
輸出 / 1M tokens	$1.2	$3.75	0.32×
快取讀取 / 1M tokens	$0.03	$0.075	0.4×
快取寫入	不額外計費	-	-

費率取自建置時的即時目錄；各模型頁面皆附有目前的費率卡。

它們的相對位置 — 在此計費單位下，所有 69 個聊天模型的每 1M tokens 的輸入價格（對數尺度）

DeepSeek V4.1 Flash · $0.3 Gemini 3.7 Flash · $0.75

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

能力

	DeepSeek V4.1 Flash	Gemini 3.7 Flash
工具使用	是	是
思考控制	是 —— 廠商未公布調節參數	是 —— 廠商未公布調節參數
結構化輸出	-	是
提示快取	隱式（自動）	隱式 + 顯式
快取生命週期	no fixed TTL (evicted when unused)	未公開
最小快取前綴	未公開	4096 個 token

規格

	DeepSeek V4.1 Flash	Gemini 3.7 Flash
輸入模態	文字影像	文字影像音訊影片
輸出模態	文字	文字
發布日期	2026-09-10	2026-08-13
知識截止日期	-	2026-03
上下文視窗	1M	1M
最大輸出	393K	66K

規格摘錄自各供應商的文件；供應商未發布的資料列會直接省略，而非自行推測。完整來源： DeepSeek V4.1 Flash · Gemini 3.7 Flash

單一提示詞，兩款模型 — 經由閘道測量

針對 CFO 的推論修復包含錯誤的函式擷取至 schema 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

DeepSeek V4.1 Flash 通過 · 3 sentences

Enabling reasoning changed the unit economics: the model now spends extra billable tokens internally to “think” before producing the answer, and those reasoning tokens are charged like output tokens even though users never see them. That added token generation—plus longer contexts and occasional retries on harder prompts—roughly doubled our cost per request. We can control this by capping reasoning depth, routing only complex queries to reasoning models, and caching common prompts to bring the bill back down.

輸出 423 tok (+327 思考) 延遲 14.6 s

Gemini 3.7 Flash 通過 · 3 sentences

Enabling reasoning causes the model to generate a high volume of invisible "thinking tokens" behind the scenes to deliberate and solve complex logic before delivering a final response. Because our vendor bills strictly on token volume, this additional internal computation effectively doubled our billable output per query. In short, our costs increased because we are now paying for the extra compute capacity required to produce significantly higher-accuracy answers.

輸出 878 tok (+799 思考) 延遲 6.9 s

指令遵循（恰好三句，可數）、受眾適配（面向 CFO 的語氣），以及下方 token 計量所暴露的隱藏思考計費缺口。

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. 檢查修復通過測試

DeepSeek V4.1 Flash 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 else: cur = 1 best = max(best, cur) return best ``` The bug is that `cur` is counting equal adjacent pairs, not elements in the current run: every time `xs[i] == xs[i - 1]`, the run length increases by one element, but the first element of the run was never counted, so all runs are undercounted by one. Also, `best` starts at `0`, so single-element inputs return `0` instead of `1`, and a one-element run ending at the last index is missed entirely.

輸出 918 tok (+733 思考) 延遲 11.4 s

Gemini 3.7 Flash 通過 · 8/8 cases

Here is the corrected function: ```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 best = max(best, cur) return best ``` ### Explanation of the Bug The original function tracked equal adjacent transitions instead of the actual number of elements in a run, causing all non-empty run lengths to be undercounted by one (e.g., returning `0` instead of `1` for single-element lists because the loop never executes, and `1` instead of `2` for pairs). Initializing both `best` and `cur` to `1` (after handling empty inputs) and resetting `cur` to `1` on a mismatch ensures every individual element is counted as a valid run of at least length one.

輸出 1050 tok (+827 思考) 延遲 6.4 s

修復是否真的正確（可執行）、解釋的資訊密度，以及在一個邊界明確的任務上的 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. 檢查合法 JSON，schema 精確

DeepSeek V4.1 Flash 通過 · 5/5 fields, guidance "withheld pending the Q4 close"

{"company":"Meridian Robotics","quarter":"September quarter","revenue_usd_m":412,"growth_pct":11.0,"guidance":"withheld pending the Q4 close"}

輸出 1707 tok (+1667 思考) 延遲 15.1 s

Gemini 3.7 Flash 通過 · 5/5 fields, guidance "withheld"

```json { "company": "Meridian Robotics", "quarter": "September quarter", "revenue_usd_m": 412, "growth_pct": null, "guidance": "withheld" } ```

輸出 918 tok (+858 思考) 延遲 6.2 s

schema 服從度（不臆造欄位）、幻覺壓力（guidance 明確被暫緩給出），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

DeepSeek V4.1 Flash 通過 · 116 words, 0 banned, 1 question

Today we're launching Cross-Provider Prompt Cache for our API gateway. It stores identical prompt requests and their responses across supported model providers, then serves cached results when a match is found. Teams can cut duplicate inference costs, reduce latency, and keep behavior consistent during provider failover. The cache works with configurable TTLs, per-route rules, and cache-key controls, so you decide what is reusable and what must stay fresh. Does your application send the same prompts to multiple providers? Now your gateway can answer many of those calls without another upstream request. Existing observability dashboards show hit rates, saved tokens, and estimated spend reduction. Enable it in the gateway console, set your policy, and start caching today.

輸出 917 tok (+770 思考) 延遲 9.0 s

Gemini 3.7 Flash 通過 · 120 words, 0 banned, 1 question

Why pay twice for identical AI queries simply because you routed them to different model vendors? Today, we introduce Universal Prompt Caching directly within our unified API gateway architecture. This capability stores repeated prompt contexts across OpenAI, Anthropic, and local models, instantly returning stored results to eliminate redundant computation fees. When your application sends an LLM request, the gateway inspects the payload, identifies semantic matches, and returns accurate cached responses in under ten milliseconds. Engineering teams can now slash inference latency by eighty percent while dramatically reducing monthly token expenditures across diverse production deployments. You retain complete privacy control, flexible cache eviction policies, and granular metrics through a single dashboard. Update your routing settings today to accelerate overall system performance.

輸出 2858 tok (+2718 思考) 延遲 14.1 s

約束服從度（字數預算、禁用詞表、唯一的那句問句）、文風指紋，以及長度控制。

只需一行程式碼即可在兩者間切換

以下每個頁籤中都有這兩個 ID — 醒目提示的這兩行是唯一的修改處。相同的端點，相同的金鑰，相同的請求結構。

from openai import OpenAI

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

resp = client.chat.completions.create(
    model="deepseek-v4.1-flash",
    # model="gemini-3.7-flash",  # 取消註解此行，並註解上一行
    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: "deepseek-v4.1-flash",
  // model: "gemini-3.7-flash",  // 取消註解此行，並註解上一行
  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": "deepseek-v4.1-flash",
    # "model": "gemini-3.7-flash",  # 取消註解此行，並註解上一行
    "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: "deepseek-v4.1-flash",
		// Model: "gemini-3.7-flash",  // 取消註解此行，並註解上一行
		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("deepseek-v4.1-flash")
        // .model("gemini-3.7-flash")  // 取消註解此行，並註解上一行
        .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

DeepSeek V4.1 Flash 和 Gemini 3.7 Flash 哪個比較便宜？

DeepSeek V4.1 Flash 在輸入 / 1m tokens 上較便宜（$0.3 對比 $0.75，相差 2.5×）。其他項目可能呈現相反結果 — 上表提供完整資訊，實際成本取決於您的使用組合。

我可以在不進行兩次整合的情況下，對 DeepSeek V4.1 Flash 和 Gemini 3.7 Flash 進行 A/B 測試嗎？

可以。兩者皆透過同一個相容 OpenAI 的端點提供服務，並使用同一把 API 金鑰 — 切換只需更改一行的模型字串，因此您可以將部分流量分別導向兩者並直接比較帳單。

DeepSeek V4.1 Flash 與 Gemini 3.7 Flash 支援提示快取嗎？

是的 — 兩者的快取讀取費率皆低於其輸入費率，因此具有暖前綴的工作負載成本會低於牌價所示。確切的快取讀取列請見上方的定價表。

DeepSeek V4.1 Flash vs Gemini 3.7 Flash

Benchmark 成績

定價

能力

規格

單一提示詞，兩款模型 — 經由閘道測量

只需一行程式碼即可在兩者間切換

常見問題

相關比較