Pricing Fresh 🌱
Complete pricing reference for Gemini Embedding 2, including free tier, standard, and batch rates.
Pricing Overview
Free Tier
| Modality | Price |
|---|---|
| Text | Free |
| Images | Free |
| Audio | Free |
| Video | Free |
| Documents | Free |
Free for All
The free tier covers all input types at no charge, making it ideal for development and testing.
Standard Pricing (Paid Tier)
| Modality | Per 1M Tokens | Per Unit |
|---|---|---|
| Text | $0.20 | - |
| Images | $0.45 | ~$0.00012 per image |
| Audio | $6.50 | ~$0.00016 per second |
| Video | $12.00 | ~$0.00079 per fps |
Batch Pricing (50% Discount)
| Modality | Per 1M Tokens | Per Unit |
|---|---|---|
| Text | $0.10 | - |
| Images | $0.225 | ~$0.00006 per image |
| Audio | $3.25 | ~$0.00008 per second |
| Video | $6.00 | ~$0.000395 per fps |
Comparison with gemini-embedding-001
| Feature | gemini-embedding-001 | Gemini Embedding 2 |
|---|---|---|
| Text (Standard) | $0.15/1M tokens | $0.20/1M tokens |
| Text (Batch) | $0.075/1M tokens | $0.10/1M tokens |
| Images | Not available | $0.45/1M tokens |
| Audio | Not available | $6.50/1M tokens |
| Video | Not available | $12.00/1M tokens |
| Free tier | Yes (text only) | Yes (all modalities) |
Cost Estimation Examples
Example 1: Text-Only RAG (1M documents)
| Scenario | Tokens | Cost (Standard) | Cost (Batch) |
|---|---|---|---|
| 1M docs, avg 500 tokens each | 500M tokens | $100.00 | $50.00 |
| Daily queries (10K/day) | ~5M tokens/day | $1.00/day | - |
| Monthly query cost | ~150M tokens | $30.00/month | - |
Example 2: Image Search (100K images)
| Scenario | Units | Cost (Standard) | Cost (Batch) |
|---|---|---|---|
| 100K images indexed | 100K images | $12.00 | $6.00 |
| 1K image queries/day | 1K queries | $0.20/day (text) | - |
Example 3: Video Library (10K clips, 60s avg)
| Scenario | Units | Cost (Standard) | Cost (Batch) |
|---|---|---|---|
| 10K videos, 60s each | 600K seconds | ~$96.00 | ~$48.00 |
| Text search queries | 10K/day | $2.00/day | - |
Cost Optimization Tips
- Use batch API for bulk operations (50% savings)
- Reduce dimensions to lower storage costs (not API costs, but downstream)
- Cache embeddings — never re-embed the same content
- Use the free tier for development and testing
- Chunk efficiently — optimal chunk sizes reduce total tokens
- Match task types — better results mean fewer re-queries
Dimension Impact on Storage (Not API Cost)
API pricing is per token regardless of output dimensions, but storage costs vary:
| Dimensions | Bytes per Vector | Storage per 1M Vectors |
|---|---|---|
| 3072 | 12,288 | ~12 GB |
| 1536 | 6,144 | ~6 GB |
| 768 | 3,072 | ~3 GB |
| 256 | 1,024 | ~1 GB |
| 128 | 512 | ~0.5 GB |
See Also
- API Reference — Full endpoint documentation
- Benchmarks — Quality vs dimension tradeoffs
- Migration — Cost comparison with v1