Reduce token usage by sending only the records, fields, and conversation history the model needs—not by assuming that shorter JSON or another format is automatically cheaper. Measure each change with the tokenizer and usage reporting for your target API, and check that the model still handles the task correctly.
1. Remove data the task does not need
Start with the task, then decide which records and fields are necessary to complete it. Filter or aggregate data in your application when the model does not need the underlying detail. Removing irrelevant input reduces what the model must process; OpenAI’s [latency optimization guidance](https://platform.openai.com/docs/guides/latency-optimization) discusses input-token reduction for especially large contexts.
For example, if the task is to identify overdue invoices, sending customer support notes, unrelated account fields, and paid invoices adds material that may not help produce the answer. Keep the data needed to identify and evaluate overdue invoices, along with enough context to interpret it.
2. Measure the request as the target model sees it
Characters, bytes, and tokens are different measures. A shorter string is not necessarily a lower-token string, and a plain-text tokenizer count may not include every token in a complete API request. OpenAI recommends tiktoken for programmatic plain-text tokenization, but advises checking API usage for the full request.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware match#1 Best Overall
- Read Before You Buy — No Video Output: These adapters support charging and USB 2.0 data transfer, but cannot transmit video signals. Except for standard USB webcams (which use USB data only), they are not compatible with HDMI/DisplayPort cables, video-capable USB-C hubs, or docking stations with video output.
- Convert USB-A Ports to USB-C: Designed to connect USB-C earphones, cables, flash drives, card readers, and other USB-C accessories to standard USB-A ports. Plug-and-play with no drivers or software required.
- Aluminum Alloy Housing: Built with a sturdy aluminum alloy shell that aids in heat dissipation and protects against daily wear and scratches. Designed to maintain a stable and secure connection.
- Compact & Travel-Friendly: The ultra-compact design allows the adapter to stay plugged into your device without blocking adjacent ports or adding bulk, reducing wear and tear on your original USB ports.
- 12-Month Warranty: Backed by a 12-month manufacturer warranty for peace of mind. Designed to meet strict quality control standards for reliable everyday performance.
Establish a baseline with representative requests. Record input-token usage, relevant usage details, task quality, latency, and cost. Use a local tokenizer estimate to iterate, then compare it with the API’s reported usage for the target model and configuration. Recheck after changing the model, tokenizer, schema, tool definitions, or API settings.
3. Choose a representation based on the data
Minifying JSON removes whitespace, but the token impact depends on the content and tokenizer. Compare minified and pretty-printed JSON on real examples rather than inferring savings from character count. Keep the representation that is both economical and clear enough for the model to interpret reliably.
Rank #2
- 5-in-1 USB-C Hub: Experience comprehensive connectivity featuring a Power Delivery input, two USB-A 2.0 ports, a USB-A 3.0 port, and an HDMI port. (Note: The USB-C power delivery input port is only for connecting an external wall charger to power your laptop and cannot power peripheral devices.)
- 90W Pass-Through Charging: Achieve optimal charging with 90W pass-through power to your laptop, supported by a total input of 100W, with the hub reserving 10W for operational efficiency. (Note: Wall charger not included.)
- Quick Data Transfers: Accelerate your productivity with rapid data transfers using a high-speed 5Gbps USB 3.0 port and two 480Mbps USB 2.0 ports.
- 4K HDMI Display: Enhance your visual experience with a hub capable of delivering 4K resolution at 30Hz in both mirror and extend modes. Please note that this hub is compatible with MacBook (macOS 12 and newer), Windows 10 and 11, ChromeOS, and laptops equipped with DP Alt Mode and Power Delivery. Note: This device is not compatible with Linux.
- What You Get: Anker USB-C Hub (5-in-1, 4K HDMI), welcome guide, 18-month warranty, and our friendly customer service.
| Representation | When to test it | Trade-off |
|---|---|---|
| JSON | Nested, irregular, or varied records | Explicit keys and structure can aid interpretation, but repeated keys may add tokens. |
| CSV or schema-once rows | Flat tables or repeated records with the same fields | Can avoid repeating labels, but requires an unambiguous column order and instructions. |
| TOON | Large, regular datasets where a schema-once layout may fit | Worth evaluating for that narrower use case; it is often a poor fit for deeply nested or non-uniform structures. CSV can be more compact for flat tables. |
Thoughtworks’ 2026 Technology Radar treats TOON as an option to assess for large, regular datasets—not a universal JSON replacement—and notes its limitations for nested or irregular data. Whatever format you test, compare token count and task quality with the target model. Do not assume a format change has a universal savings percentage.
4. Reduce repeated material and make compact data interpretable
Repeated keys, labels, and instructions can add input tokens. When records share a stable shape, test defining the field order once and sending compact rows. For example, an instruction can state that each row contains invoice_id, due_date, amount, status, followed by rows in that order.
Rank #3
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
Only abbreviate labels or omit repeated structure when the model can still interpret every value without guessing. Include the schema or mapping the model needs, and validate the output against representative cases. A smaller prompt that makes fields ambiguous may save tokens while undermining the task.
5. Reuse stable prefixes for recurring requests
If many requests share the same instructions or schema, put that stable material before changing user-specific data. OpenAI says, “Prompt caching reuses work when requests share the same prompt prefix.” Its guide says cached-input rates can be discounted “up to 95%,” but that is not a guaranteed per-request saving: eligibility, model, cache mode, pricing, and the amount of reusable input affect the result. Check the current Prompt Caching guide and inspect cached-token usage rather than assuming a cache hit.
Rank #4
- Dual Converters, Infinite Potential：Includes 2× USB C male to USB A female adapters and 2× USB A male to USB C female adapters. Perfect for a wide range of uses—tablets with Bluetooth keyboards, expand USB ports on macbook, and more. Two different converters for all your daily needs
- Next-Level 10Gbps & 3A Charging: No more slow 480Mbps, this usb to usb c adapter has a transfer speed of up to 10Gbps, allowing you to do more transferring in less time. This usb adapter fits both USB A and USB C charger, supporting up to 3A fast charging
- Upgraded Exquisite Craftsmanship: With an aluminum alloy housing and metal connector, the usbc to usb adapter is extremely durable and sturdy. Rigorously tested to withstand more than 10,000 times of plugging and unplugging, ensuring long-lasting performance
- Broad Compatible: The usb c to usb adapter widely supports all USB C/ USB A devices like laptops, tablets, cellphones, car chargers, and phone chargers. Such as compatible with MacBook Pro/Air 2023/2022, Thunderbolt 4/3 Devices,Apple MagSafe Watch 9/8/7/SE/Ultra, iPad Pro 2022/2021, Samsung Galaxy S23/S20/S10, and iPhone 17/16/15 Pro. Plug and play
- Please Note: To reach 10Gbps speed, keep the cable under 3.3 ft. For USB A Male to USB C adapters, try flipping the USB C connector. USB C Male to USB A adapters support bidirectional 10Gbps transfer within 3.3 ft
Keep the reusable prefix identical where practical, and place volatile records or user-specific values afterward. Caching applies to a matching eligible prefix; it does not eliminate the need to process new input. Before enabling extended cache retention, review the provider’s current data-control documentation for your chosen configuration.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.6. Compact accumulated conversation history carefully
In a long-running interaction, earlier turns can consume context even when their details are no longer useful. Remove obsolete history or use a supported compaction flow that preserves the state needed for the next task. OpenAI documents server-side compaction for reducing context size while retaining state in long interactions.
Free tools Windows power users keep installed
One-click scans. No signup required.
Best Value
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
After compaction, evaluate total usage and task quality, not just the apparent length of the compacted text. The changed prompt may alter prefix matching and cache reuse. See OpenAI’s compaction documentation for the supported flow and behavior.
7. Keep input and output optimization separate
Structured-output instructions and schemas can add input tokens, while the requested output format affects output tokens. If the goal is to reduce what you send, measure the input side separately from output formatting. A shorter output representation may help in suitable cases, but it does not by itself reduce the structured data already sent in the request. OpenAI’s Structured Outputs documentation explains the format feature; evaluate its input and output effects in your own usage data.
Quick Recap
A practical test sequence
- Choose representative tasks and capture a baseline: target model and settings, API-reported input usage, quality, latency, and cost.
- Remove unnecessary records and fields, then rerun the same tasks.
- Compare minified JSON, the original representation, and a schema-once format if the records are regular enough to support it.
- For repeated calls, move stable instructions and schema before changing data, then verify cached usage in the API response.
- For long conversations, test trimming or supported compaction and compare total usage, cache behavior, and task quality.
- Keep a change only when it improves the relevant cost or context measure without making results less reliable.
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.
