Imagen 4 vs Imagen 3
Benchmarks, API Token Pricing, Context Window & Speed Comparison
Related topics (12)
Imagen 4 by Google and Imagen 3 by Google represent two of the frontier AI models analyzed in our comprehensive 2026 AI benchmark and LLM comparison. In this head-to-head evaluation, we examine API token pricing, context window specifications, inference speed (tokens per second), and empirical reasoning performance. From a pure economic perspective, Imagen 3 holds a notable pricing advantage for token ingestion. On output generation, costs are matched. In data-heavy architectures requiring extensive repository or document analysis, Imagen 4 provides an expanded memory horizon with 2K tokens. In latency-sensitive production environments, Imagen 4 leads in tokens-per-second throughput, making it ideal for real-time conversational agents and high-concurrency workflows. Below is the complete empirical head-to-head breakdown across cost, speed, context, and capability matrices.
任意のAIモデルを切り替えて比較
Select any AI models from our verified benchmark database to generate live side-by-side matrices.
Imagen 4
フラッグシップ最先端モデルGoogle が提供する Imagen 4 は、最大 2K トークンのコンテキストウィンドウを備え、入力 $40/100万トークン・出力 $0/100万トークンで利用可能な、コーディング・推論・業務自動化向け AI モデルです。
$40.00 / 1M
$0.00 / 1M
2K tokens
N/A
Imagen 3
フラッグシップ最先端モデルGoogle が提供する Imagen 3 は、最大 1K トークンのコンテキストウィンドウを備え、入力 $30/100万トークン・出力 $0/100万トークンで利用可能な、コーディング・推論・業務自動化向け AI モデルです。
$30.00 / 1M
$0.00 / 1M
1K tokens
N/A
ひと目でわかる主な違い
- Imagen 3 is 1.3x cheaper for API usage (input + output tokens combined).
- Imagen 4 has a larger context window (2K vs 1K tokens).
- Imagen 4 is faster at 25 tok/s vs 20 tok/s.
| Feature | Imagen 4 | Imagen 3 |
|---|---|---|
Intelligence Index Composite reasoning & knowledge score (0-100) | 82 | 72 |
Provider | ||
Input Price Cost per million input tokens | $40.00 / 1M | $30.00 / 1M |
Output Price Cost per million output tokens | $0.00 / 1M | $0.00 / 1M |
Blended Price Weighted 7:2:1 cache:input:output per 1M | $8.00 / 1M | $6.00 / 1M |
Context Window Maximum input + output tokens | 2K tokens~3 A4 pages of text | 1K tokens~2 A4 pages of text |
Max Output Tokens Per single request | 4K | 4K |
Throughput Speed Tokens generated per second | 25 tok/s | 20 tok/s |
SWE-bench Verified Real-world software engineering | N/A | N/A |
分野別勝者スコアカード
At-a-glance decision frameworkEvaluations ongoing
検証済みベンチマークスコア比較
Official evaluation metrics verified by independent third-party research labs.
Crowdsourced human preference evaluation based on blind, head-to-head model battles.
Tokens emitted per second during active generation streaming.
レイテンシ＆応答時間
Time to first token (TTFT) and estimated end-to-end response time for a typical 500-token response.
リアルタイムAPIトークン料金シミュレーター
Adjust sliders to calculate projected monthly cloud API spending with prompt caching discounts.
Imagen 4 · Strengths & Key Advantages
- Google の高度な High-resolution multimodal image diffusion transformer アーキテクチャと 2K トークンのコンテキスト処理能力
- 優れたコストパフォーマンス（入力 $40/1M、出力 $0/1M）と安定した応答速度
- プログラミング、論理的推論、多言語タスクにおいて高いベンチマーク精度を実現
Considerations & Trade-offs
- 大規模な長文コンテキスト処理時はトークン消費量の最適化が必要
- 専門領域のタスクでは適切なプロンプト設計や RAG の併用が推奨される
Imagen 3 · Strengths & Key Advantages
- Google の高度な Latent text-to-image diffusion transformer アーキテクチャと 1K トークンのコンテキスト処理能力
- 優れたコストパフォーマンス（入力 $30/1M、出力 $0/1M）と安定した応答速度
- プログラミング、論理的推論、多言語タスクにおいて高いベンチマーク精度を実現
Considerations & Trade-offs
- 大規模な長文コンテキスト処理時はトークン消費量の最適化が必要
- 専門領域のタスクでは適切なプロンプト設計や RAG の併用が推奨される
Token Use & Task Economics
Estimated cost and token consumption to complete one representative agentic task.
| Metric | Imagen 4 | Imagen 3 |
|---|---|---|
Cost per Task Estimated USD to finish one task | $0.04 | $0.03 |
Output Tokens / Task Tokens generated per task | 0K | 0K |
Reasoning Tokens / Task Thinking tokens burned per task | 0K | 0K |
技術仕様・スペック
Architecture, release timeline, capabilities, and modalities.
| Feature | Imagen 4 | Imagen 3 |
|---|---|---|
| Official API Model ID (SDK) | models/imagen-4 | models/imagen-3 |
| API Protocol | Google Gemini API (generateContent) + OpenAI-Compatible SDK | Google Gemini API (generateContent) + OpenAI-Compatible SDK |
| Release Date | 2025-05 | 2024-08 |
| Knowledge Cutoff | March 2025 | July 2024 |
| Architecture | High-resolution multimodal image diffusion transformer | Latent text-to-image diffusion transformer |
| Parameters | Confidential | Confidential |
| Enterprise Rate Limits (RPM / TPM) | 10,000 – 30,000 RPM (Tier 5 Enterprise) · 800,000 – 2,000,000 TPM | 10,000 – 30,000 RPM (Tier 5 Enterprise) · 800,000 – 2,000,000 TPM |
| Long-Context Tier (>128K/200K) | Flat rate ($40/1M in) across full 2K window | Flat rate ($30/1M in) across full 1K window |
| Prompt Cache Policy & TTL | Implicit + Explicit Context Caching at $0/1M with configurable TTL | Implicit + Explicit Context Caching at $0/1M with configurable TTL |
| Computer Use / GUI Agent | Text-only completion (No GUI action loop) | Text-only completion (No GUI action loop) |
| MCP & Parallel Tool Calling | Prompt-based JSON extraction only | Prompt-based JSON extraction only |
| Web Search Grounding & Sandbox | Built-in Google Search Grounding + Managed Python Code Execution Sandbox | Built-in Google Search Grounding + Managed Python Code Execution Sandbox |
| Official Fine-Tuning Support | Vertex AI Supervised Adapter Tuning & Context Distillation | Vertex AI Supervised Adapter Tuning & Context Distillation |
| Multimodal Token Billing | Text BPE (~4 chars/token) | Text BPE (~4 chars/token) |
| Quantization & Min GPU VRAM | Cloud-Hosted Proprietary (FP8 / BF16 Tensor Core Cluster) · Zero Local VRAM Required (100% Serverless Cloud API) | Cloud-Hosted Proprietary (FP8 / BF16 Tensor Core Cluster) · Zero Local VRAM Required (100% Serverless Cloud API) |
| Safety & Frontier Classification | Google DeepMind Frontier Safety Framework Verified | Google DeepMind Frontier Safety Framework Verified |
| Enterprise Compliance (ZDR/HIPAA/SOC2) | SOC2 Type II · GDPR · Zero Data Retention (ZDR) & HIPAA BAA Available on Enterprise Tier | SOC2 Type II · GDPR · Zero Data Retention (ZDR) & HIPAA BAA Available on Enterprise Tier |
| Input Modalities | text | text |
| Output Modalities | image | image |
Imagen 4 vs Imagen 3: よくある質問 (FAQ)
Which is more intelligent, Imagen 4 or Imagen 3?
Imagen 4 is more intelligent. It scores 82 compared with 72 on the composite Intelligence Index.
Which model is faster?
Imagen 4 is faster, generating 25 tokens per second compared with 20 tokens per second.
Which model is cheaper to run?
Imagen 3 is cheaper at $6.00 per 1M tokens blended, compared with $8.00 (7:2:1 cache:input:output ratio).
Which model has lower latency?
Imagen 4 has lower latency, with a time to first token of 2.00s compared with 2.50s.
Which model has a larger context window?
Imagen 4 has a larger context window, supporting 2K tokens compared with 1K tokens.
総合評価
Verdict & Model Selection Recommendation: Choose Imagen 4 if your priority is Professional advertising, print-ready media, and studio graphic design. Its performance in demanding tasks justifies its operational token cost. Choose Imagen 3 if your workflow prioritizes Commercial asset generation, marketing collateral, and creative art direction. For large-scale batch processing or customer-facing conversational interfaces, its balance of speed and economics delivers superior ROI.
よくある質問 (FAQ)
How does Imagen 4 compare to Imagen 3 in API token pricing and cost per token?+
Imagen 4 costs $40.00 per 1M input tokens and $0.00 per 1M output tokens, whereas Imagen 3 costs $30.00 input and $0.00 output. Overall, Imagen 3 is the more cost-effective model for high-volume API pipelines.
Which model has a larger context window: Imagen 4 or Imagen 3?+
Imagen 4 supports a context window of 2K tokens (~2k words), while Imagen 3 supports 1K tokens (~1k words). Therefore, Imagen 4 can process more textual data and document context in a single prompt.
Which AI model is better for software engineering and coding: Imagen 4 or Imagen 3?+
Imagen 4 scores at the frontier of coding benchmarks, while Imagen 3 scores competitively for software generation. For complex autonomous debugging and agentic engineering, Imagen 4 provides robust code generation and tool calling.
What is the difference in reasoning capability and extended thinking between Imagen 4 and Imagen 3?+
Imagen 4 is optimized for low-latency direct generation. In contrast, Imagen 3 focuses on standard fast inference.
How does inference speed (tokens per second) compare between Imagen 4 and Imagen 3?+
Imagen 4 generates approximately 25 tokens per second, while Imagen 3 delivers 20 tokens per second. In latency-sensitive production environments, Imagen 4 provides faster time-to-first-token (TTFT) and higher streaming response velocity.
Can you implement a tiered LLM routing setup combining Imagen 4 and Imagen 3 to optimize costs?+
Yes! High-efficiency architectures route high-frequency routine classification and conversational tasks to Imagen 3 to save operational costs, and dynamically escalate complex architectural queries or reasoning tasks to Imagen 4. This tiered routing pattern cuts overall API bills by up to 70%.
Understand key metrics like context window, token pricing, reasoning mode, and SWE-bench scores in our comprehensive AI glossary.
Related reading
- GPT Image 1 vs Imagen 4: Benchmarks, Pricing & Speed Comparison比較
- Imagen 3 vs GPT Image 1: Benchmarks, Pricing & Speed Comparison比較
- Nano Banana vs Imagen 4: Benchmarks, Pricing & Speed Comparison比較
- Aion-3.0 vs Aion-Mini: Benchmarks, Pricing & Speed Comparison比較
- Aion-3.0 vs Gemini 3.6 Flash: Benchmarks, Pricing & Speed Comparison比較
- Aion-Mini vs Laguna XS 2.1: Benchmarks, Pricing & Speed Comparison比較
Benchmark scores, token prices, and rate limits evolve rapidly as AI labs release model updates. Always verify current specifications and pricing on the official provider documentation.
公的ソース: GetInfoUs Independent AI Benchmarks & Official API Specs