Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Not on the figures this firm reported. The Call Center Doctors rented four Nvidia H200 GPUs to test DeepSeek V4.1 Flash, but found that running the model on hardware at the regular rental rate did not beat its estimated DeepSeek API cost for the workload. Nor did the trial establish a like-for-like 80-times saving over Claude: the headline comparison mixes token list prices with the company’s much lower Claude Code subscription bill. A security problem also kept DeepSeek from writing code, so this was a test of read-only code review—not a head-to-head coding-agent benchmark.
What the company tested—and what it did not
The Call Center Doctors says it rented a four-H200 server after eight-GPU systems were unavailable. It ran DeepSeek V4.1 Flash in a short experiment on September 27, 2026. The company reports five starts before the model was stable, with each load taking roughly 10–15 minutes. Its spot instance was reclaimed within minutes of the final test.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
MINISFORUM MS-02 Ultra Workstation Mini PC, Intel Core Ultra 9 285HX (24C/24T, up to 5.5GHz), PCIe... | $1,659.00 | Buy on Amazon |
| 2 |
|
GMKtec EVO-X2 AI Mini PC Ryzen Al Max+ 395 Superchip 128GB LPDDR5X 2TB SSD | $3,649.99 | Buy on Amazon |
As an Amazon Associate I earn from qualifying purchases.
The consultancy’s account is the only primary source located for the trial’s workload logs and cost estimates; Tom’s Hardware reported on that account but did not reproduce the test. The figures below are therefore company-reported measurements or estimates unless identified otherwise—not independently replicated results.
Security findings changed the scope
During its setup review, the firm says it found sandbox escape paths, including a settings file in a shared temporary folder that could let agent code run as administrator. It does not provide a technical exploit write-up or independent security review, and the account does not establish that the issue applies to other agent environments.
#1 Best Overall
- High-Performance AI Processor：The MS-02 Ultra features an Intel Core Ultra 9 285HX (24C/24T, up to 5.5 GHz, 13 TOPS NPU), delivering fast and efficient performance for AI inference, algorithm development, and media workloads. A PCIe x16 expansion slot supports desktop-class GPU upgrades for advanced model training and accelerated computing tasks. It's ideal for creators, engineers, and teams handling intensive parallel workloads.
- 4 × M.2 PCIe 4.0 + 4 × DDR5 SODIMM slots:Four DDR5 SODIMM slots support up to 256 GB of memory, while ECC helps maintain data integrity in mission-critical environments. Four PCIe 4.0 M.2 slots support up to 24 TB of storage, supporting RAID 0/1/5/10, combining high-speed performance with data protection. It allows for the creation of independent scratch disks, media libraries, and project drives, providing high-throughput for production workflows.
- PCIe & USB 4.0 v2: Up to three PCIe slots can be equipped, including a dual-slot x16 GPU. The main slot supports PCIe 5.0, meeting the needs of high-bandwidth creative and computing workloads. USB 4.0 v2 (80Gbps) supports high-bandwidth external storage and displays.
- Ultra-fast Networking: Wi-Fi 7 further enhances wireless performance with next-generation speeds and low-latency stability. Intelligent bandwidth switching optimizes throughput in different network environments, ensuring optimal performance for enterprise or local networks. Dual 25GbE ports (providing up to approximately 3.125 GB/s bandwidth, about 25 times faster than traditional 1GbE), enabling seamless large-scale file transfers and parallel computing. 10GbE and 2.5GbE ports, with support for Intel vPro technology, ensure enterprise-grade remote management and deployment flexibility.
- Server-grade thermal architecture: Utilizing a dedicated CPU/GPU airflow design, equipped with a 6-pipe dual-fan cooler, it maintains stable performance even under sustained loads, delivering up to 140W Turbo power while maintaining a 100W TDP, and operating with noise levels as low as 36 dB. An integrated 350W power supply ensures stable and reliable output for demanding computing tasks and fully loaded extended configurations.
Because of that finding, the company kept its code-writing agents offline. DeepSeek served only read-only reviewer agents; the firm says it shipped zero lines of DeepSeek-written code during the test. The results cannot show whether DeepSeek could complete code changes as safely or effectively as Claude in a comparable builder workflow.
Why “80x cheaper” is not the same as a cheaper bill
The 80x framing concerns token list prices, not the amount this firm actually paid for Claude Code. Those are different cost bases. A per-token API price, a subscription bill, and the cost of keeping rented GPUs running cannot be treated as interchangeable.
Three different ways to pay
| Cost basis | What the company reported | What the figure means |
|---|---|---|
| Rented hardware | $18.37 per hour at the regular rate; $9.19 per hour for spot, according to The Call Center Doctors (2026) | The regular instance costs money continuously while rented, whether busy or idle. The spot instance was cheaper but could be reclaimed by the provider. |
| DeepSeek API | Estimated $184–$223 per day for this workload at full utilization, according to The Call Center Doctors (2026) | A workload-based estimate, not a universal price for other customers. It avoids paying for an always-on rented server. |
| Claude Code subscription | About $5,500 for September 1–27, according to The Call Center Doctors (2026) | The firm’s reported subscription spend for that period—not a token-list-price calculation. |
At $18.37 an hour, a full 24-hour day costs $440.88, as the company calculated. That is roughly two to 2.4 times its estimated $184–$223 daily DeepSeek API cost for the same work. Spot rental at $9.19 an hour would cost about half as much as regular rental, but it is not an equivalent always-available service because the provider could reclaim the instance.
Tom’s Hardware calculated that the regular rate would amount to about $13,200 for a standard-length month. That arithmetic is useful for understanding the hardware bill, but it is not a same-period comparison with the firm’s approximately $5,500 Claude subscription spend from September 1–27. The amounts suggest why the title’s “doubles Claude bill” shorthand needs qualification: one is an extrapolated month of continuously rented hardware, the other is a reported 27-day subscription bill.
Reported token prices are not the firm’s effective costs
The company’s first-party account lists DeepSeek off-peak rates of $0.15 per million new-input tokens, $0.003 per million cached-input tokens and $0.60 per million output tokens; weekday peak rates were reported as double. It lists Claude Opus 5.5 rates of $4 per million input tokens, $20 per million output tokens and $0.20 per million cache reads. These are the rates as reported in the account, not a guarantee of current pricing.
The consultancy estimates that its September 1–27 usage would have cost $3,500 at DeepSeek API rates off-peak, $7,000 at peak, or about $4,200 if usage were evenly spread across the period. It also says that pricing Claude Opus 5.5’s same token volume at list rates would imply about $140,000. That last figure is not what it paid: its reported Claude Code subscription spend was about $5,500. A list-price token calculation cannot be presented as a direct comparison with the subscription bill.
Rank #2
- EVOLUTION 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 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, 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; 12% 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.
For September, the company reports 5,610 merged changes. It says Claude subscriptions worked out to about $1 per change, while its DeepSeek estimate was $1.15–$4.90 per change after allowing for more tokens, lower success and Claude checking. The latter is explicitly an estimate, not a result from DeepSeek completing code changes in this trial, since its builder agents stayed offline.
The workload mattered as much as the GPU count
The firm’s isolated speed tests measured different kinds of token processing under one-minute full-load runs. It reported 16,621 tokens per second for reading new text, 521,027 per second for rereading cached text and 5,281 per second for writing. Its long-answer test reached 5,871 written tokens per second. These figures should not be read as end-to-end coding-agent throughput: the company says agent conversations repeatedly resubmit prior context, creating a different mix of cached reads, new reads and output.
In its September workload, the firm says 96% of model input was rereading prior conversation. Its stated mix was 41.6 new tokens and 1,042 old cached tokens read per token written. On that mix, it calculated that four H200s could produce about 213 written tokens per second, or roughly 20 billion total tokens a day. It compared that with 51 billion tokens on its busiest September day. The consultancy says its formula matched its live test within 3%; no independent replication was identified.
This is why raw token-per-second figures are not enough to decide whether self-hosting makes sense. A buyer needs to account for how much context is new versus cached, how much the agent writes, and how often conversations are resubmitted. A workload dominated by cached context can behave very differently from a test that reads only new input or generates long answers.
What the read-only review trial showed
The consultancy says it operated 48–64 read-only reviewer agents. They read 2,377 folders and filed 32 bug reports. Those activity counts describe the review use it actually ran; they do not establish the quality of DeepSeek-generated code, the rate of successfully completed coding tasks, or a safe sandbox configuration for other organizations.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
The company’s reported workload comparison points to a practical decision rule: compare the service outcome, not just a model’s listed token rates. Include subscription or API charges, GPU hours even when idle, setup and operations time, spot-instance interruptions, the cached/uncached context mix, successful changes, review work, and the isolation controls needed to run generated code. This trial did not settle the broader question of whether self-hosting can win under a different workload or safer environment.
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.
