2 ms·
From opencode go $10/mo plan I get between 60 t/s and 100 token/s even with large contexts of 150k+ tokens. I wouldn't call 80 t/s slow.
by bel8 2mo ago
From opencode go $10/mo plan I get between 60 t/s and 100 token/s even with large contexts of 150k+ tokens.
I wouldn't call 80 t/s slow.
- ponyous 2mo agoYou are right, relatively to other llm providers this is not slow. But if you think what is possible when you have 1000t/s a sec you might find it slow.
- hatefulmoron 2mo agoThat's across 64 concurrent streams; you could make more concurrent requests to DeepSeek API no?