🚀 A OpenWebUI function to log your context and completions to Weave (by Weights & Biases) for LLMOps / observability.
Get started tracking your @OpenWebUI context and completions in 90 seconds or less.
github.com/m-rgba/openwebui-…
246
Martin McFly retweeted
Just updated my pricing sheet for hosted LLMs. It includes the latest from @OpenAI, @AnthropicAI, @GoogleDeepMind, @MistralAI, @AIatMeta, @cohere, @awscloud Bedrock, @Microsoft Azure, @deepseek_ai, @anyscalecompute, and @togethercompute. 👀
> ~$15 per 1M output tokens from the best closed LLMs
> Deepseek v2 models cheapest at $0.28/M output tokens
> Gemini 1.5 Flash offers the best cost performance from closed providers.
> Gemini 1.5 Pro is half the price of GPT-4o, Claude Sonnet 3.5
> Llama 3 70B costs ~$1 per 1 million token
Which provider am I missing?
And lastly some great advice on building product in the AI with @sqs and the awesome title "the AI emperor has no DAUs" . promptpanel.com/expand/deep-…
3
269
Martin McFly retweeted
satoshi? oh you mean the time-traveling ai that hit us with an economic infohazard in order to accelerate gpu production?
It's 2022, I'm ready to dust off my Bjango Actions again github.com/bjango/Bjango-Act…
1





