The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →AI models do not have to be huge or expensive by definition, IBM CEO Arvind Krishna argued at the company’s Think 2025 conference in Boston. His proposal is to use smaller models tailored to specific business tasks alongside larger general-purpose systems—not to replace every large model. Whether that approach is better for a particular business depends on how well a model performs the work and what capabilities, cost, speed and deployment options it requires.
What did Arvind Krishna say about AI model size?
“There is no law of computer science that says that AI must remain expensive and must remain large,” Krishna said at IBM Think 2025. ITPro reported the remark on May 7, 2025, and it also appears in CRN’s event transcript. ITPro’s report and CRN’s transcript record his case for smaller, purpose-built models.
Krishna argued that a model designed for a particular business use case and tuned to relevant enterprise data can be faster and less costly to run, while offering more flexibility about where it is deployed. IBM’s Granite family is one example of the company’s investment in that approach. Those benefits are IBM’s and Krishna’s claims; the cited event coverage does not independently establish that smaller models are generally more accurate or cheaper in practice.
Smaller models are meant to complement larger ones
Krishna described smaller models as an addition to the AI toolkit, not a universal substitute for larger systems: “It’s not a substitute for the larger models. It’s an ‘and’ with the larger models.” A business might choose a compact model for a well-defined task and use a larger general-purpose model where the work calls for broader capabilities. Which option makes sense depends on the specific task and the model’s demonstrated performance.
#1 Best Overall
- Built for Local AI Development: AMD Ryzen AI Halo is designed for local AI development and inference, featuring 128GB unified memory and support for up to 200B parameter models to build and run intensive AI workloads locally.
- 128GB Unified Memory: Features 128GB LPDDR5x unified memory at 8000 MT/s with 256 GB/s memory bandwidth, providing a shared memory pool across the CPU, GPU, and NPU to support larger AI models.
- AMD Ryzen AI Max+ 395 Processor: Features 16 cores, 32 threads, and Zen 5 architecture, paired with AMD Radeon 8060S integrated graphics featuring 40 RDNA 3.5 compute units and an AMD XDNA 2 NPU with up to 50 TOPS.
- Linux AI Developer Platform: Purpose-built for Linux-based AI development with full AMD ROCm software support and preloaded tools, models, and workflows optimized for local AI development.
- Compact, Connected Design: Includes a 2TB M.2 SSD, 10GbE LAN, Wi-Fi 7, Bluetooth 5.4, USB-C connectivity, and HDMI 2.1b.
At the keynote, Krishna contrasted models in the 3-, 8-, 13- and 20-billion-parameter range with models in the 300- or 500-billion range. These were illustrative figures from his remarks, not a survey of the current market or a direct comparison of particular models.
Why model size alone does not settle the question
Smaller models can reduce resource demands, but parameter count by itself does not tell a business whether a model is suitable. Efficiency techniques also affect what resources a model needs. In its analysis of DeepSeek, IBM explains that Multi-Head Latent Attention reduces the size of the key-value (KV) cache, helping lower memory use. The same IBM article discusses remaining compute barriers and trade-offs, including weaker function-calling capability and safety-alignment concerns.
Those examples illustrate why an efficient model is not automatically the right model. For each candidate, check its results on the intended work, along with the operating constraints and capabilities the application depends on:
Rank #2
- EVOLUTION AMD RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 64GB pool, which is perfect for running LLMs such as Deepseek 32B, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 4% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
- Task performance: Test the model on representative examples from the business task rather than assuming size predicts quality.
- Running cost and speed: Compare inference cost and latency under the expected workload.
- Compute and memory: Consider what the model and its efficiency techniques require in the intended environment.
- Deployment and data constraints: Check whether the model can run where the organization needs it and work with the relevant data under its governance requirements.
- Required behavior: Evaluate function calling, safety and other capabilities that matter to the application.
At Think 2025, Krishna also said that “99% of all enterprise data has been untouched by AI.” That is his keynote assertion, as recorded by CRN, not an independently verified industry statistic in the cited coverage. He connected the opportunity to bringing AI to business data, but the figure should not be treated as a neutral measurement.
What the accuracy claim does—and does not—show
Krishna said smaller models are “now more accurate than larger models.” The available CRN transcript does not specify the benchmark, tasks, model versions or evaluation method behind that comparison. It therefore records his claim, but does not establish that small models are generally more accurate. To decide between models, a business needs evidence tied to its own task and requirements.
Krishna framed the shift as a move from experimentation toward practical deployment: “The era of AI experimentation is over. Success is going to be defined by integration and business outcomes.” That is the strategic case for matching models to work that organizations actually need done—not proof that a particular model size will deliver a particular result.
Rank #3
- Intel Core Ultra 9 285 Processor: Newly developed cores deliver ultra-smooth and responsive gameplay. AI accelerators prepare users for the next era of gaming on an AI PC.
- Simplistic Design: Enjoy the latest generation of Windows 11 Home for your everyday needs. *MSI recommends Windows 11 Pro for business use.
- NVIDIA GeForce RTX 5070 Ti GPU
- Cool While Gaming: In conjunction with an RGB CPU Air Cooler, the Aegis RS features four system cooling fans; three in the front and one in the rear to pull in cool air and push heat out of the PC.
- Turn on the Bright Lights: With the built-in RGB lighting, take your gaming experience to the next level by pressing the MSI LED button to cycle through lighting options. Customize lighting even further with MSI Center software.
What IBM’s Think 2025 message means for businesses
IBM’s argument is that businesses should not assume every AI use case requires the largest available general-purpose model. A smaller model may be worth evaluating when the task is narrow, the organization can test it on relevant data, and its cost, speed and deployment characteristics suit the application. Larger models remain part of the picture when their broader capabilities better fit the work.
The practical decision is not “small or large” in the abstract. It is whether a candidate model meets the application’s performance, cost, latency, infrastructure, deployment, data-governance and capability requirements. Krishna’s keynote makes the case for considering more options; it does not establish a universal winner.
Free tools Windows power users keep installed
One-click scans. No signup required.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
