You are correct, at least based on my experimentation. This subagent workflow really only delivers meaningful results when working across model families (ex: Opus asking Astra for a review) or when working across conceptual hardware boundaries (cloud offloading to local inference, etc). The planner / overseer model ends up reading all the output of the coder model anyway, at least if you want quality output, so there's not much benefit otherwise.
> The planner / overseer model ends up reading all the output of the coder model anyway
I mean, even that's fine because output tokens cost ~5x input tokens. So that's still a win if the agent that does the heavy code reading and writing produces output that may be read again by the lead (not my observation of how agents in Paseo interact and of course, correctable to degrees with shared system prompt).
Here's a snippet of a sample prompt downwards from the lead for a piece of code I'm working on:
From agent: 3ff73f80-9713-458c-8271-d275f03609c6
<role>
You are the Principal Engineer on <snipped ... /> (.NET 11 / Vue MCP server; read CLAUDE.md in your working directory first). ...
</role>
<lead_role>
I am the Director of Engineering. The System Architect and the DMTS reviewed this approach ...
</lead_role>
<sniped ... />
<comms>
Run all gating commands in the FOREGROUND. End your turn ONLY with the final report: pass/fail per success criterion (a)-(i) with evidence, the branch tip SHA and `git log --oneline origin/main..HEAD`, lines added/deleted, the exact 0factor image digest and version you pinned, every deviation from this brief and why, risks you saw,...
</comms>
The agent is given explicit "comms" instructions upwards and I find that both agents in this exchange stick to the contract.
> even that's fine because output tokens cost ~5x input tokens.
That's assuming you come in under a certain threshold of revisions, and excludes the tokens generated for all the exchanges. The later can be be significant, especially when there are significant issues to correct.
> The planner / overseer model ends up reading all the output of the coder model anyway, at least if you want quality output, so there's not much benefit otherwise.
Welcome to why the RLM paper does not explore recursive depths beyond 1. The game of context telephone as you ascend the call stack quickly falls apart.
Nah; I can use my token budget across both providers this way, use tokens more sparingly with smaller models for heavy text in/out (why pay Opus prices to babysit a PR? run tests?), and you'd be surprised when the agents talk to each other and "role play" how much this improves things like the UX as well as technical design which are more subjective.
Distinguished Member of Technical Staff, for example, is asked to "expand the solution space" which I find has a strong effect on approaches considered (it doesn't know what was already considered on the main thread; only the domain and problem space).
This also helps with context growth since the sub-agents do the heavy lifting (e.g. search, read docs) and just provide the key bits back to the lead.
Like you, I was a single-model, single-TUI user up until three weeks ago when I picked up Paseo and it is the only tool that I've found that has significantly positively improved my throughput without sacrificing quality of product.
Is it "larping"? I ran an eval today on a contract ingest and extract agent I'm working on. Changing the <role> fragment of the prompt lifted the results by +7 points by simply changing the text to the effect of "You are a senior partner at a prestigious law firm that focuses on reading licensing contracts..." and some bits about the content the agent is reviewing + objective. No one would believe it without an eval.
Claude docs explicitly say to LARP[0]:
> Give Claude a role
>
> Setting a role in the system prompt focuses Claude's behavior and tone for your use case. Even a single sentence makes a difference
My conclusion: the agents like to LARP so I'm happy to LARP with them.
Why stop at "You are a senior partner at a prestigious law firm that focuses on reading licensing contracts..." why not "You are super programmer/developer that has been programming with Ada using punchcard, having Linus on quick dial, you can smell BS code from miles away and know exactly which question to ask".
I hate to be pedantic, but LARP stands for live action role playing.
Only mentioning it because LLMs are already taking over software development and everything online. Let's keep them out of our meet space peculiarities (1) while we can. ;)
> I hate to be pedantic, but LARP stands for live action role playing.
> Only mentioning it because LLMs are already taking over software development and everything online.
In our glorious future, you'll have your agent LARP for you with your friends' agents, so you all can spend more time doing more fulfilling tasks like watching ads on Instagram.
Yeah at the point that Opus has read enough context to hand off a detailed plan, implementing the plan may not be that many more tokens.
Also having spent the weekend with the supposedly great Deepseek Flash 4.1 1017, I wonder if the people singing its praises have used recent frontier models. DS Flash feels like maybe Sonnet 4.5-4.6. Absolutely falls apart at things Opus/Sol currently excel at with low reasoning, and even has basic problems with failing tool calls, which hasnt been an issue with Anthropic/OpenAI in forever.
Could be a configuration issue with the provider I tried but honestly, I didnt get much better out of GLM5.3-flash. GLM5.3 and Kimi3 are much better but since most inference providers cant take a loss they are more expensive than just using Opus/Sol on subs, which are almost certainly currently a loss leader for their respective API offeringss.
No "may be" about it. There is no benefit to having bad models write your code, and there's no benefit to "plan" mode. Plan mode is an outdated 2025 concept that should be removed from standard harnesses.
There are 2 scenarios for dumb builder agents:
1. Your "planner" actually implements the project almost completely, finds all the gotchas, solves all of the emergent problems, and fully plans the solution.
2. Your "planner" takes a half-assed first attempt, solving none of the major problems that will arise during implementation
In scenario 1 -- just pay cache output on a few more tokens to finish the project. No need to spend twenty minutes pumping a few million tokens through deepseek regardless of how cheap it is to do so.
In scenario 2 -- good luck getting quality output
I do think this "software team" LARPing is an engineering version of AI psychosis. You can get all of the little parts humming along, sending messages, and looking really cool. You can even convince yourself that all of these busy little bees actually make better product.
But a frontier model working alone can produce better output in a tiny fraction of the time. Just build the thing.
An obvious downside to plan mode is that without access to the edit tool, the model can't confirm its hallucinations.
Removing edit tool from your model is like lobotomizing your engineer.
Here's the secret: Just building the thing is the modern "plan mode". You don't have to keep what it made. But a completed, verified, and tested output is always more valuable to my planning process than a hallucinated sloppy guess.
My experience has been that building the thing is not modern "plan mode".
I do not use an explicit "plan mode" that tries to limit what the LLM can do. But I do tell it to explore and research and produce an artifact of what needs to happen to finish a large ticket and then use that as input to the actual implementation. I do have it even write small experiments to verify some-parts of the total design to connect it to reality.
The reason is that there can often be lots of decisions that need a large amount of context about the code base in order to reasonably determine what the right design and implementation is and this doesn't happen (in my experience) with the little Interview dialogue that pops up when it isn't sure. Or, even worse, Opus will just make decisions for you that are incoherent with the goals of the codebase.
So, I think, "plan mode" as an explicit mode is dead, but the act of planning is still a good thing to do.
Try it. 3 weeks ago I was in the same boat. Single model, TUI. I only tried Paseo primarily to test multi-host orchestration (after giving up on Herdr) and not even multi-agent, multi-plat behavior.
Output quality, especially subjective bits, has improved relative to my attention needed. Autonomous decision making is much, much better for me adding one profile that specifically is instructed to focus on UI/UX, one for end user proxy, and one for MTS.
It costs you nothing to try except maybe 15 minutes of setup. Give it a solid, skeptical week. It's good even without the profiles, but way better with.
Yes, have done cross plat plan review. But why it works so well with Paseo is that it's automatic and multi-turn (agents kind of "discuss" (I say "kind of" because agents don't really discuss like you or I would with short, multi-turn speak-listen-pivot-repeat).
Work tokens are severely limited to the extent that if I used Sonnet 5.5 for implementation, I'd be out 2h into a session (Team 5x). +$400/mo. at API rates on OpenAI. So Haiku + Luna it is. There is no other way this is even usable for agentic dev at work. Works fine for people who multi-turn with a single agent.
Personal accounts I run the $17 Anthropic plan + $100 Codex Pro + DeepSeek. It's insane how much this stretches with this setup.
Maybe the crux here is how hands on you are with your sessions. As I've moved more hands off, I'm relying on other agents to take my place by embedding my heuristics into their config.
I've had an instruction to review the code with "claude -p" in place around the time GPT-5.5 released. I think it has served me well.
If I recall correctly it was before subagents were a thing. Even back then it was already automatic and would on its own request follow-up reviews or arbitrate a plan file if asked.
I tried it and figured out the benefit was getting the agent to look at solving the problem from different perspectives. Now I just include those perspectives in the prompt to a single agent.
Example: If I feel like the agent is going to just bolt on a feature without refactoring the core parts, I tell it to expand functionality of the core parts as needed.
In my experience Opus 5.5 is good but Opus very big picture plan mode to Qwen 3.8 code is so massively massively cheaper (I'm on API) that there's no competition. This is as of a week ago.
I don't see how that applies here. In that case it's 2+ potentially compromised things working in tandem to propagate the malicious effect. In this case I'd think someone reverse engineering a compiler to see if it's compromised would ensure their tools are clean enough (use known clean sources or build their own). And if there's no way to ensure a minimal cleanness then the world is already irrecoverably lost.
So it's not something for me, but like everything Microsoft makes, it's for our BOFH's who will torment us with yet another corporate compliance machine.
A lot of us aren't actually using enough tokens to reach the value add of a $20USD/month subscription versus just paying for the API. Without even mentioning the chinese competitors, you can get a lot of tokens for $10 of ChatGPT Luna, which is more than capable of solving the kinds of problems I have at home (where I use hermes harness).
Exactly, sometimes when reading Hacker News I get the impression this thing is a cult. Luckily there's still some dissident voice around to keep it interesting
reply