I cut my AI API costs 99% by switching from Claude to DeepSeek
1–10 of 23 posts
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#2[flagged]
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#3Which models are we talking about? Is there any degradation in quality, long context retrieval?
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#4Which models are we talking about? Is there any degradation in quality, long context retrieval?
The tweet mentioned deepseek V4 flash.
From HF: 284B parameters (13B active), 1M context window.
This is indeed some kind of compressed context and the quality goes down as the context grows. IIRC the V4 paper had some numbers on this
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#5Which models are we talking about? Is there any degradation in quality, long context retrieval?
[flagged]
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#6Which models are we talking about? Is there any degradation in quality, long context retrieval?
[flagged]
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#7Which models are we talking about? Is there any degradation in quality, long context retrieval?
The tweet mentioned deepseek V4 flash. From HF: 284B parameters (13B active), 1M context window. This is indeed some kind of compressed context and the quality goes down as the context grows. IIRC the V4 paper had some numbers on this https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash
V4 flash is much worse than any Claude model. If you're doing something simple, it can be a good way to save money though.
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#8Which models are we talking about? Is there any degradation in quality, long context retrieval?
It depends on how mature the DeepSeek model became before OpenAI noticed that they were wholesale replicating their model and starting blocking access
https://www.reuters.com/world/china/openai-accuses-deepseek-...
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#9Earlier quoted context omitted.
The tweet mentioned deepseek V4 flash. From HF: 284B parameters (13B active), 1M context window. This is indeed some kind of compressed context and the quality goes down as the context grows. IIRC the V4 paper had some numbers on this https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash
V4 flash is much worse than any Claude model. If you're doing something simple, it can be a good way to save money though.
[flagged]
Re: I cut my AI API costs 99% by switching from Claude to DeepSeek
#10Earlier quoted context omitted.
The tweet mentioned deepseek V4 flash. From HF: 284B parameters (13B active), 1M context window. This is indeed some kind of compressed context and the quality goes down as the context grows. IIRC the V4 paper had some numbers on this https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash
V4 flash is much worse than any Claude model. If you're doing something simple, it can be a good way to save money though.
I agree that Claude is better (definitely better than the flash version which is relatively small). But...
I actually canceled my Claude Code plan a few months back after trying out some of the "lesser" models on openrouter. They seem to work as just as well (or just as bad) for my coding tasks.