I'm considering the possibility that it's good to break the prefix and cache because the LLM itself was rewarded (during post-training) with different prefixes/system prompts, each containing reasoning traces of the correct size.
I might be very very wrong though and LLMs disagree with me, insisting that cache is preserved and the system message doesn't have to change (even though it often contains effort level in context) if effort level changes across turns, and that all you have to do is tell the inference lib that parses think tags to early-close think tags that are too long.
The number of tokens you predict at time (multi or not) has nothing to do with whether the model wants to emit any, some or a lot of reasoning tokens in reasoning tag -- similar to how branch prediction will not really change the for loop iteration count.
Take a look at the harmony repo which specifies the internal OpenAI format - the effort level is specified in the context after the <|start|> tag - https://github.com/openai/harmony
Note that inference libs also have parsers that put hard limits on reasoning tokens with separate counters (similar to how you can put a limit on token generation per completion versus waiting for an <eos>). For that, take a look at vllm reasoning docs.
Sounds believable to me. Google is really good at data warehouse & database tech (BigQuery, BigTable, Spanner) and access management systems, as well as sturdy custom-built tools for internal use by employees.
Even though it might not be intentional, Apple is lagging at such tooling.
I might be very very wrong though and LLMs disagree with me, insisting that cache is preserved and the system message doesn't have to change (even though it often contains effort level in context) if effort level changes across turns, and that all you have to do is tell the inference lib that parses think tags to early-close think tags that are too long.