Loading page
From $0.1163 per 1M input tokens and $0.4651 per 1M output tokens today, platform fee included, on Gondola's OpenAI- and Anthropic-compatible API.
| Unit | Gondola today, fee included | DeepSeek API verified 2026-10-18 | Venice list |
|---|---|---|---|
| Input, per 1M tokens | $0.1163 | $0.300 |
| $0.3000 |
| Output, per 1M tokens | $0.4651 | $1.20 | $1.2000 |
|---|
| Cached input, per 1M tokens | $0.0029 | not on file | not shown |
|---|
| Cache write, per 1M tokens | no cache tier | not on file | not shown |
|---|
At today's cheapest listed offer that is 61.2% below DeepSeek's own API price, for a typical message of 300 input plus 400 output tokens.
Cache-miss input, PEAK tier (off-peak is half: $0.15/$0.60; peak hours 01:00-04:00 and 06:00-10:00 UTC, Mon-Fri), matching the tier the sibling deepseek-v4-pro-0813 entry uses. Sold under the FLOATING id deepseek-flash, whose pricing-table MODEL VERSION row names this build (DeepSeek's 2026-10-17 changelog: V4.1 Flash upgraded to DeepSeek-V4.1-Flash-1017, pricing unchanged). The 1:1 claim rests on that row plus Venice's id: it carries the 1017 build date, was listed 2026-10-18 as a new row beside the unchanged undated one (Venice's pattern for every dated DeepSeek build, -0731 and -0813 included) at the same prices and 1M context. The row's name and modelSource are copied from the undated row (huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash, where the dated repo is DeepSeek-V4.1-Flash-1017), so they say nothing either way; re-read if Venice renames or repoints the row, or when DeepSeek repoints deepseek-flash again.
Gondola today is the cheapest offer listed right now, including the platform fee of up to 2 percent. It moves as sellers join, reprice, or drain. The maker's and Venice's figures are their posted prices for a direct customer, before any fee of their own.
curl https://api.gondola-ai.com/v1/chat/completions \
-H "Authorization: Bearer $GONDOLA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v4-1-flash-1017",
"messages": [
{
"role": "user",
"content": "Hello"
}
]
}'