Extraction estimate · prices checked 2026-11-03
Gemini 3.8 Flash vs Levanto Sage 2 for field extraction
List prices
| Model | $ per M input | $ per M output | Source |
|---|---|---|---|
| Gemini 3.8 Flash | $0.75 | $3.75 | Google pricing |
| Gemini 3.1 Flash-Lite | $0.25 | $1.5 | Google pricing |
| Levanto Sage 2 | $0.05 | $1 | Levanto list price |
Checked 2026-11-03. Same table the public prices endpoint serves. Gemini 3.8 Flash: introductory price through Dec 31, 2026, then $1.50 / $7.50. At the later price Gemini 3.8 Flash would cost about $4.20 per 1,000 documents and the break-even share drops to 45%.
What we measured and what we assumed
- Measured: Levanto Sage 2 used 329 input and 607 output tokens per document on average, over 36 synthetic lead, ticket and invoice documents (228 field values), in our bench run on 2026-11-03. Its output includes per-field probabilities, which is why it is larger than the JSON alone.
- Assumed for Gemini 3.8 Flash: 700 input tokens (a ~300-token document plus the audit’s default 400-token prompt with field definitions) and 120 visible output tokens for the JSON.
- Our estimate: 300 reasoning tokens at Gemini 3.8 Flash’s API default effort (
medium), billed as output. Extraction gets 2.5× our yes/no estimate.
That gives an estimated $2.10 per 1,000 documents on Gemini 3.8 Flash, against $0.624 on Levanto Sage 2.
How accurate was Levanto Sage 2 on our bench?
| Set | Documents | Field values right | Lead / ticket / invoice |
|---|---|---|---|
| Clean templates | 36 | 96.9% of 228 | 98.8% / 90% / 100% |
| Messier wording | 15 | 88.4% of 95 | 82.9% / 88% / 94.3% |
Weakest fields: ticket.amount_paid (58% on clean documents) and lead.budget (20% on the messier set). Money amounts and free-form budgets are where a fallback to Gemini 3.8 Flash matters. Levanto Sage 2 left fields empty correctly 84.1% of the time when they were absent. This is a small synthetic bench, not a benchmark of your data.
Effort level changes the bill
Estimated Gemini 3.8 Flash cost per 1,000 documents by effort level, with reasoning billed as output:
| Effort | Est. reasoning tokens | Per 1,000 documents |
|---|---|---|
| low | 100 | $1.35 |
| medium (API default) | 300 | $2.10 |
| high | 800 | $3.98 |
If the fields are copied straight from the text, a lower effort may give the same JSON; test it on your documents. When lowering effort beats switching models.
Monthly estimate by volume
“Share kept” is a scenario: the share of documents where Levanto Sage 2 is confident on every field. The rest also go to Gemini 3.8 Flash.
| Documents / month | All on Gemini 3.8 Flash | Hybrid, 50% kept | 70% kept | 90% kept |
|---|---|---|---|---|
| 10,000 | $21.00 | $16.74 | $12.54 | $8.34 |
| 100,000 | $210 | $167 | $125 | $83.36 |
| 1,000,000 | $2,100 | $1,674 | $1,254 | $834 |
Estimates at the assumptions above. The 30% bar needs Levanto Sage 2 to keep at least 60% of documents. For yes/no calls on Gemini 3.8 Flash, see Gemini 3.8 Flash vs Levanto Sage 2 for classification.
Check it on your own documents
Paste 5 to 25 documents with the JSON Gemini 3.8 Flash returned and your field definitions into the free cutover audit. It reports agreement per field, picks the confidence threshold and projects cost with your token counts. How it works · Other models · Is my task too simple for a big model?