LLM System Design and Mannequin Choice – O’Reilly

Choosing the proper LLM has turn into a full-time job. New fashions seem virtually every day, every providing totally different capabilities, costs, and quirks, from reasoning strengths to value effectivity to code era. This competitors creates sturdy incentives for AI labs to carve out a distinct segment and offers new startups room to emerge, leading to a fragmented panorama the place one mannequin might excel at reasoning, one other at code, and a 3rd at value effectivity.

AI, in a single sense, is getting cheaper quicker than any earlier know-how, at the least per unit of intelligence. For instance, enter tokens for Gemini 2.5 Flash-Lite are roughly 600 occasions cheaper than what OpenAI’s GPT-3 (davinci-002) value in August 2022, whereas outperforming it on each metric. On the similar time, entry to frontier capabilities can be changing into dearer than ever. The reason being easy: we will now pay straight for extra functionality, which has led to the rise of $300+ per thirty days Professional subscription tiers.

In the present day, any developer can run succesful open-weight fashions domestically for negligible marginal value utilizing instruments like Ollama. On the similar time, enterprise methods can expertise sharp value will increase, relying on the mannequin measurement (variety of parameters, resembling 3 billion, 70 billion, and even within the trillions), the variety of inner processing steps, and the amount of enter knowledge. For builders, these are central system design decisions that straight have an effect on feasibility and value construction. For finish customers, this complexity explains why a primary subscription differs a lot from a premium plan with increased limits on superior fashions.

The alternatives you make in these broader improvement choices additionally decide which LLM and inference settings are optimum on your use case.

At In direction of AI, we work throughout the LLM stack, constructing functions, designing enterprise methods, and providing on-line programs (together with one on O’Reilly), customized company coaching, and LLM improvement consultancy. In our expertise, mannequin choice and system design have turn into central to getting significant outcomes from these instruments. A lot of that, in flip, will depend on the place right this moment’s fashions are gaining their capabilities. Whereas scale nonetheless performs a task, latest progress has come from a broader combine of things, together with training-data high quality, post-training strategies, and particularly how fashions are used at inference time.

The Shifting Foundations of Mannequin Functionality

Whereas early positive aspects in LLM efficiency tracked intently with will increase in pretraining compute, bigger datasets, larger fashions, and extra coaching steps, this strategy now yields diminishing returns.

Latest enhancements come from a broader mixture of methods. Pretraining-data high quality has turn into simply as essential as amount, with higher filtering and AI-generated artificial knowledge contributing to stronger fashions. Architectural effectivity, just like the improvements launched by DeepSeek, has began to shut the hole between measurement and functionality. And post-training strategies, particularly instruction tuning and reinforcement studying from human or AI suggestions (RLHF/RLAIF), have made fashions extra aligned, controllable, and responsive in apply.

The extra basic shift, nonetheless, is occurring at inference time. Since late 2024, with fashions like OpenAI’s o1, we’ve entered a brand new part the place fashions can commerce compute for reasoning on demand. Fairly than relying solely on what was baked in throughout coaching, they’ll now “suppose more durable” at runtime, operating extra inner steps, exploring different solutions, or chaining ideas earlier than responding. This opens up new functionality ceilings, but in addition introduces new value dynamics.

These different enchancment methods have led to a transparent divergence amongst AI labs and fashions, a speedy growth in mannequin selection, and in some instances, an explosion in mannequin utilization prices.

The Trendy Value Explosion: How Inference Scaling Modified the Sport

Inference-time compute scaling has launched a brand new dynamic in LLM system design: We’ve gone from a single lever mannequin measurement, to at the least 4 distinct methods to commerce value for functionality at runtime. The result’s a widening hole in inference value throughout fashions and use instances, typically by components of 10,000x or extra.

Bigger fashions (measurement scaling): The obvious lever is sheer mannequin measurement. Frontier LLMs, like GPT-4.5, typically constructed with combination of consultants (MoE) architectures, can have enter token prices 750 occasions increased than streamlined fashions like Gemini Flash-Lite. Bigger parameter counts imply extra compute per token, particularly when a number of consultants are lively per question.

Sequence scaling (“pondering tokens”): Newer “reasoning” LLMs carry out extra inner computational steps, or an extended chain of thought, earlier than producing their closing reply. For instance, OpenAI’s o1 used ~30x extra compute than GPT-4o on common, and sometimes 5x extra output tokens per activity. Agentic methods introduce a further technique of collection scaling and an additional layer of value multiplication. As these brokers suppose, plan, act, reassess, plan, act, and so forth, they typically make many LLM steps in a loop, every incurring further value.

Parallel scaling: Right here, the system runs a number of mannequin cases on the identical activity after which robotically selects one of the best output through automated strategies, resembling majority voting (which assumes the most typical reply is probably going right) or self-confidence scores (the place the mannequin output claiming the best confidence in its response is taken as one of the best). The o3-pro mannequin doubtless runs 5–10x parallel cases over o3. This multiplies the price by the variety of parallel makes an attempt (with some nuance).

Enter context scaling: In RAG pipelines, the variety of retrieved chunks and their measurement straight affect enter token prices and the LLM’s capability to synthesize an excellent reply. Extra context can typically enhance outcomes, however this comes at a better value and potential latency. Context isn’t free; it’s one other dimension of scaling that builders should price range for.

Taken collectively, these 4 components characterize a basic shift in how mannequin value scales. For builders designing methods for high-value issues, 10,000x to 1,000,000x variations in API prices to unravel an issue primarily based on architectural decisions at the moment are practical potentialities. Reasoning LLMs, though solely distinguished for about 9 months, reversed the development of declining entry prices to the perfect fashions. This transforms the choice from “Which LLM ought to I exploit?” to incorporate “How a lot reasoning do I need to pay for?”

This shift adjustments how we take into consideration choice. Selecting an LLM is not about chasing the best benchmark rating; it’s about discovering the steadiness level the place functionality, latency, and value align along with your use case.

Core Mannequin Choice Standards

When selecting a mannequin we discover it is very important first clearly determine your use case and the minimal core AI capabilities and attributes wanted to ship it.

A typical first step is to try commonplace benchmark scores (for instance LiveBench, MMLU-Professional, SWE-Bench). These benchmarks are a helpful start line, however some fashions are tuned on benchmark knowledge, and real-world efficiency on duties which might be truly related to you’ll typically fluctuate. Filtering benchmark assessments and scores by your business and activity class is a priceless step right here. An LLM optimized for software program improvement would possibly carry out poorly in artistic writing or vice versa. The match between a mannequin’s coaching focus and your utility area can outweigh general-purpose benchmarks.

Leaderboards like LMArena and Synthetic Evaluation provide broader human‑choice comparisons however nonetheless don’t exchange customized real-world testing. It helps to have a set of your individual instance questions or duties at hand to check out a brand new mannequin for your self and see the way it performs. This could embrace a mixture of simple duties to determine a baseline and difficult edge instances the place it’s simple for a mannequin to make errors.

As you progress past advert hoc testing, for any severe improvement effort, customized evaluations are non-negotiable. They have to be tailor-made to your use case and the varieties of issues you remedy. That is the one solution to really know if a mannequin, or a change to your system, is genuinely enhancing issues for your customers and your particular enterprise objectives.

Listed below are some core components we take into account:

Multimodality is rising as a serious differentiator. Fashions like GPT-4o and Gemini can deal with not simply textual content but in addition photos, audio, and in some instances video, unlocking functions that pure textual content fashions can’t help.

Context window and efficient context window utilization are additionally key: What number of tokens or paperwork can the mannequin course of and the way a lot of that marketed context window can the LLM truly use successfully with out efficiency degradation relative to duties that use much less context?

Latency is very vital for interactive functions. Normally, smaller or cheaper fashions have a tendency to reply quicker, whereas reasoning-heavy fashions introduce delays as a consequence of deeper inner computation.

Reasoning is the power to scale inference-time compute and carry out multistep problem-solving, planning, or deep evaluation.

Privateness and safety are sometimes key concerns right here. For instance, if you wish to hold your mental property non-public, you need to use a mannequin that received’t practice in your inputs, which frequently factors towards self-hosted or particular enterprise-grade API options.

Trustworthiness can be changing into essential and may come all the way down to the repute and monitor document of the AI lab. A mannequin that produces erratic, biased, or reputationally damaging outputs is a legal responsibility, no matter its benchmark scores. As an illustration, Grok has had well-publicized points with its alignment. Even when such points are supposedly mounted, it creates a lingering query of belief: How can one make certain it received’t behave equally sooner or later?

Moreover, the data cutoff date additionally issues whether it is for use in a fast-moving discipline.

After understanding if a mannequin meets your minimal functionality, the following resolution is usually on optimizing trade-offs amongst value, reliability, safety, and latency. A key rule of thumb we discover helpful right here: If the reliability achieve from a dearer mannequin or extra inference time saves extra of your or your customers’ time (valued when it comes to pay) than the mannequin prices, going with the bigger mannequin is an effective resolution!

The Execs and Cons of Open-Weight and Closed-API LLMs

The rise of more and more aggressive open-weight LLMs, resembling Meta’s Llama collection, Mistral, DeepSeek, Gemma, Qwen, and now OpenAI’s GPT-OSS has added a vital dimension to the mannequin choice panorama. Momentum behind this open ecosystem surged with the discharge of DeepSeek’s R1 reasoning mannequin, aggressive with OpenAI’s o1 however priced at roughly 30x decrease API prices. This sparked debate round effectivity versus scale and intensified the broader AI rivalry between China and the US. Reactions ranged from “OpenAI and Nvidia are out of date” to “DeepSeek’s prices have to be fabricated,” however no matter hype, the discharge was a milestone. It confirmed that architectural innovation, not simply scale, may ship frontier-level efficiency with far higher value effectivity.

This open-model offensive has continued with sturdy contributions from different Chinese language labs like Alibaba (Qwen), Kimi, and Tencent (Hunyuan), and has put aggressive stress on Meta after its open-weight Llama fashions fell behind. China’s latest management in open-weight LLMs has raised new safety/IP points with some US- and European-based organizations, although we observe accessing these mannequin weights and operating the mannequin by yourself infrastructure doesn’t require sending knowledge to China.

This brings us again to the professionals and cons of open weights. Whereas closed-API LLMs nonetheless lead on the frontier of functionality, the first benefit of open-weight fashions is fast and inexpensive native testing, unparalleled flexibility, and elevated knowledge safety when run internally. Organizations may carry out full fine-tuning, adapting the mannequin’s core weights and behaviors to their particular area, language, and duties. Open fashions additionally present stability and predictability—you management the model you deploy, insulating your manufacturing methods from surprising adjustments or degradations that may typically happen with unannounced updates to proprietary API-based fashions.

Public closed-model APIs from main suppliers profit from immense economies of scale and extremely optimized GPU utilization by batching requests from 1000’s of customers, an effectivity that’s tough for a single group to duplicate. This typically implies that utilizing a closed-source API could be cheaper per inference than self-hosting an open mannequin. Safety and compliance are additionally extra nuanced than they first seem. Whereas some organizations should use self-hosted fashions to simplify compliance with rules like GDPR by holding knowledge fully inside their very own perimeter, this locations your complete burden of securing the infrastructure on the interior workforce—a posh and costly endeavor. Prime API suppliers additionally typically provide devoted cases, non-public cloud endpoints, and contractual agreements that may assure knowledge residency, zero-logging, and meet stringent regulatory requirements. The selection, due to this fact, just isn’t a easy open-versus-closed binary.

The boundary between open and closed fashions can be changing into more and more blurred. Open-weight fashions are more and more provided through API by third-party LLM inference platforms, combining the flexibleness of open fashions with the simplicity of hosted entry. This hybrid strategy typically strikes a sensible steadiness between management and operational complexity.

Main Closed LLMs

Beneath, we current some key prices and metrics for main closed-source fashions accessible through API. Many of those fashions have further complexity and different pricing together with choices for quick modes, pondering modes, context caching, and longer context.

We current the newest LiveBench benchmark rating for every mannequin as one measure for comparability. LiveBench is a repeatedly up to date benchmark designed to supply a “contamination-free” analysis of huge language fashions by frequently releasing new questions with goal, verifiable solutions. It scores fashions out of 100 on a various set of difficult duties, with a major deal with capabilities like reasoning, coding, and knowledge evaluation. The same LiveBench scores between GPT-4.5 and a pair of.5 Flash-Lite, regardless of 750x enter token value variation, highlights each that smaller fashions at the moment are very succesful but in addition that not all capabilities are captured in a single benchmark!

Main open-weight LLMs

Beneath, we additionally current key prices, the LiveBench benchmark rating, and context size for main open-weight fashions accessible through API. We evaluate hosted variations of those fashions for simple comparability. Completely different API suppliers might select to host open-weight fashions with totally different ranges of quantization, totally different context lengths, and totally different pricing, so efficiency can fluctuate between suppliers.

Whether or not hosted or self-deployed, deciding on a mannequin solely solves a part of the issue. In apply, a lot of the complexity and alternative lies in how that mannequin is used: the way it’s prompted, prolonged, fine-tuned, or embedded inside a broader workflow. These system-level choices typically have a higher affect on efficiency and value than the mannequin selection itself.

A Sensible Information to Designing an LLM System

Merely choosing the largest or latest LLM is never the optimum technique. A simpler strategy begins with a deep understanding of the developer’s toolkit: understanding which method to use to which drawback to realize the specified functionality and reliability with out pointless value. That is all a part of the fixed “march of nines” as you develop LLM methods modularly to unravel for extra reliability and functionality. There’s a must prioritize the simplest wins that ship tangible worth earlier than investing in additional incremental and sometimes expensive accuracy enhancements. The truth will at all times fluctuate on a case-by-case foundation, however here’s a fast information to navigating this course of.

Step 1: Open Versus Closed?

That is typically your first resolution.

• Go together with a closed-API mannequin (e.g., from OpenAI, Google, Anthropic) if: Your precedence is accessing absolutely the state-of-the-art fashions with most simplicity.
• Go together with an open-weight mannequin (e.g., Llama, Mistral, Qwen, DeepSeek) if:
  • Knowledge safety and compliance are paramount: If it’s essential assure that delicate knowledge by no means leaves your individual infrastructure.
  • You want deep customization and management: In case your purpose is to fine-tune a mannequin on proprietary knowledge and to create a specialised professional that you simply management utterly.

In case you went open, what are you able to realistically run? Your personal GPU infrastructure is a tough constraint. Assess your cluster measurement and reminiscence to find out in the event you can effectively run a big, main 1 trillion+ parameter MoE mannequin, resembling Kimi K2, or in case you are higher served by a medium-size mannequin resembling Gemma 3 27B or a a lot smaller mannequin like Gemma 3n that may even run on cell.

Step 2: Gauging the Want for Reasoning

Does your activity require the mannequin to easily blast out a response, or does it must suppose first?

• Reasoning: For duties that contain complicated, multistep problem-solving, brainstorming, strategic planning, intricate code era, or deep evaluation, you want a devoted reasoning mannequin resembling o3, Gemini 2.5 Professional, DeepSeek R1, or Claude 4. In some instances these fashions can be utilized in high-reasoning mode, which inspires the mannequin to suppose for longer earlier than responding.
• No reasoning: For simple duties like easy Q&A, summarization of a single doc, knowledge extraction, or classification, a robust reasoning mannequin is overkill.
• The center floor: For duties requiring reasonable reasoning, resembling producing a structured report from just a few knowledge factors or performing primary knowledge evaluation at scale, a “mini” reasoning mannequin, like OpenAI’s o4-mini or Gemini Flash 2.5, affords a steadiness of functionality and value.

Step 3: Pinpointing Key Mannequin Attributes

Past basic intelligence and reasoning, fashionable LLMs are specialists. Your selection ought to be guided by the precise attributes and “superpowers” your utility wants.

• Prioritize accuracy over value for high-value duties the place errors are expensive or the place a human professional’s time is being saved. o3-pro is a standout mannequin right here and it might even be used as a reality checker to meticulously verify the small print of an earlier LLM output.
• Prioritize pace and value over accuracy: For user-facing, real-time functions like chatbots or high-volume, low-value duties like easy knowledge categorization, latency and value are paramount. Select a hyper-efficient “flash” or “mini” mannequin resembling Gemini 2.5 Flash-Lite. Qwen3-235B fashions can be an amazing choice right here however are too complicated to inference your self.
• Do you want a deep, long-context researcher? For duties that require synthesizing data from large paperwork, total codebases, or in depth authorized contracts, a mannequin with an unlimited and extremely efficient context window is essential. Gemini 2.5 Professional excels right here.
• Is multimodality important? In case your utility wants to know or generate photos, course of audio in actual time, or analyze video, your selection narrows to fashions like GPT-4o or the Gemini household. For one-shot YouTube video processing, Gemini is the standout.
• Is it a code-specific activity? Whereas many fashions can code, some are explicitly tuned for it. Within the open world, Codestral and Gemma do an honest job. However Claude has received hearts and minds, at the least for now.
• Do you want dwell, agentic net search? For answering questions on present occasions or matters past the mannequin’s data cutoff, take into account a mannequin with a built-in, dependable net search, resembling o3.
• Do you want complicated dialogue and emotional nuance? GPT-4.5, Kimi K2, Claude Opus 4.0, or Grok 4 do an amazing job.

Step 4: Prompting, Then RAG, Then Analysis

Earlier than you dive into extra complicated and expensive improvement, at all times see how far you will get with the only strategies. This can be a path of escalating complexity. Mannequin selection for RAG pipelines is usually centered round latency for finish customers, however lately extra complicated agentic RAG workflows or long-context RAG duties require reasoning fashions or longer context capabilities.

1. Immediate engineering first: Your first step is at all times to maximise the mannequin’s inherent capabilities by means of clear, well-structured prompting. Usually, a greater immediate with a extra succesful mannequin is all you want.
2. Transfer to retrieval-augmented era (RAG): In case your mannequin’s limitation is an absence of particular, non-public, or up-to-date data, RAG is the following logical step. That is one of the best strategy for decreasing hallucinations, offering solutions primarily based on proprietary paperwork, and making certain responses are present. Nonetheless, RAG just isn’t a panacea. Its effectiveness is fully depending on the standard and freshness of your dataset, and constructing a retrieval system that constantly finds and makes use of the most related data is a major engineering problem. RAG additionally comes with many related choices, resembling the amount of knowledge to retrieve and feed into the mannequin’s context window, and simply how a lot use you make of long-context capabilities and context caching.
3. Iterate with superior RAG: To push efficiency, you will want to implement extra superior strategies like hybrid search (combining key phrase and vector search), re-ranking retrieved outcomes for relevance, and question transformation.
4. Construct customized analysis: Guarantee iterations in your system design, additions of latest superior RAG strategies, or updates to the newest mannequin are at all times transferring progress ahead in your key metrics!

Step 5: Fantastic-Tune or Distill for Deep Specialization

If the mannequin’s core conduct—not its data—remains to be the issue, then it’s time to think about fine-tuning. Fantastic-tuning is a major endeavor that requires a high-quality dataset, engineering effort, and computational assets. Nonetheless, it might allow a smaller, cheaper open-weight mannequin to outperform a large generalist mannequin on a selected, slim activity, making it a robust software for optimization and specialization.

• Fantastic-tuning is for altering conduct, not including data. Use it to show a mannequin a selected talent, fashion, or format. For instance:
  • To reliably output knowledge in a posh, structured format like particular JSON or XML schemas.
  • To grasp the distinctive vocabulary and nuances of a extremely specialised area (e.g., authorized, medical).
  • Some closed-source fashions can be found for fine-tuning through API resembling Gemini 2.5 Flash and varied OpenAI fashions. Bigger fashions are usually not accessible.
  • In open-weight fashions, Llama 3.3 70B and Qwen 70B are fine-tuning staples. The method is extra complicated to fine-tune an open-weight mannequin your self.
• Mannequin distillation may function a production-focused optimization step. In its easiest type, this consists of producing artificial knowledge from bigger fashions to create fine-tuning datasets to enhance the capabilities of smaller fashions.
• Reinforcement fine-tuning (RFT) for problem-solving accuracy
  As an alternative of simply imitating right solutions, the mannequin learns by trial, error, and correction. It’s rewarded for getting solutions proper and penalized for getting them flawed.
  • Use RFT to: Create a real “professional mannequin” that excels at complicated duties with objectively right outcomes.
  • The benefit: RFT is extremely data-efficient, typically requiring just a few dozen high-quality examples to realize important efficiency positive aspects.
  • The catch: RFT requires a dependable, automated “grader” to supply the reward sign. Designing this grader is a vital engineering problem.

Step 6: Orchestrated Workflows Versus Autonomous Brokers

The vital resolution right here is how a lot freedom to grant. Autonomous brokers are additionally extra prone to want dearer reasoning fashions with higher ranges of inference scaling. Parallel inference scaling strategies with a number of brokers are additionally starting to ship nice outcomes. Small errors can accumulate and multiply throughout many successive agentic steps so the funding in a stronger extra succesful mannequin could make all of the distinction in constructing a usable product.

• Select an orchestrated workflow for predictable duties
  You design a selected, typically linear, sequence of steps, and the LLM acts as a robust element at a number of of these steps.
  • Use when: You might be automating a identified, repeatable enterprise course of (e.g., processing a buyer help ticket, producing a month-to-month monetary abstract). The purpose is reliability, predictability, and management.
  • Profit: You keep full management over the method, making certain consistency and managing prices successfully as a result of the quantity and kind of LLM calls are predefined.
• Construct hybrid pipelines: Usually, one of the best outcomes will come from combining many LLMs, open and closed, inside a pipeline.
  • This implies utilizing totally different LLMs for various phases of a workflow: a quick, low-cost LLM for preliminary question routing; a specialised LLM for a selected subtask; a robust reasoning LLM for complicated planning; and maybe one other LLM for verification or refinement.
  • At In direction of AI, we regularly have 2-3 totally different LLMs from totally different corporations in an LLM pipeline.
• Select an autonomous agent for open-ended issues. You give the LLM a high-level purpose, a set of instruments (e.g., APIs, databases, code interpreters), and the autonomy to determine the steps to realize that purpose.
  • Use when: The trail to the answer is unknown and requires dynamic problem-solving, exploration, or analysis (e.g., debugging a posh software program challenge, performing deep market evaluation, planning a multistage challenge).
  • The vital threat—runaway prices: An agent that will get caught in a loop, makes poor choices, or explores inefficient paths can quickly accumulate monumental API prices. Implementing strict guardrails is vital:
    • Price range limits: Set laborious caps on the price per activity.
    • Step counters: Restrict the whole variety of “ideas” or “actions” an agent can take.
    • Human-in-the-loop: Require human approval for doubtlessly costly or irreversible actions.
  • Gemini 2.5 Professional and o3 are our favorite closed-API fashions for agent pipelines, whereas in open-weight fashions we like Kimi K2.

Working by means of these steps helps translate a obscure drawback right into a concrete implementation plan, one which’s grounded in clear trade-offs and tailor-made to your wants. This structured strategy typically yields methods that aren’t solely extra succesful and dependable but in addition far simpler for particular duties than a general-purpose chatbot ever might be.

Conclusion

The open-versus-closed race provides us speedy entry to sturdy LLMs but in addition creates complexity. Choosing and deploying them calls for each engineering self-discipline and financial readability.

Growing within the LLM ecosystem calls for a brand new degree of engineering self-discipline and eager financial consciousness. No single LLM is a cure-all. A sensible, evolving toolkit is crucial, however understanding which software to tug out for which job is the actual artwork. The problem isn’t simply choosing a mannequin from an inventory; it’s about architecting an answer. This requires a scientific strategy, transferring from high-level strategic choices about knowledge and safety all the way down to the granular, technical decisions of improvement and implementation.

The success of specialised “LLM wrapper” functions like Anyscale/Cursor for coding or Perplexity for search, a few of which at the moment are valued at over $10 billion, underscores the immense worth on this tailor-made strategy. These functions aren’t simply skinny wrappers; they’re subtle methods that leverage basis LLMs however add important worth by means of customized workflows, fine-tuning, knowledge integration, and person expertise design.

In the end, success hinges on knowledgeable pragmatism. Builders and organizations want a pointy understanding of their drawback area and a agency grasp of how value scales throughout mannequin selection, collection and parallel reasoning, context utilization, and agentic conduct. Above all, customized analysis is non-negotiable as a result of your use case, not a benchmark, is the one commonplace that actually issues.