AI AGENT COMPUTE BENCHMARK # Coding PC Ladder How much faster is a better computer for Codex, Claude Code and other AI coding tools? Mainstream Macs, laptops and desktops ranked and tiered, with minimum, recommended and best configurations. Updated weekly. Data updated: 2026-11-02 00:00 UTC REQUIREMENTS ## RequirementsTiered on a heavy “front-end build / large project” workload; reference machine MacBook Air M4 / Mac mini M4. Memory tiers are our recommendation ### Best 15%+ faster than the reference machine; spending more adds little Mac MacBook Pro M5 Max 18-core, MacBook Pro M5 Pro 18-core, MacBook Pro M5 Pro 15-core, Mac Studio M3 Ultra 28-core Windows laptop — Desktop Desktop Core Ultra 9 285K, Desktop Core Ultra 7 265K, Desktop R9 9950X, Desktop i9-14900K, Desktop i7-14700K Memory 32 GB or more ### Recommended Matches or beats the reference machine; comfortable for daily and heavy work Mac MacBook Pro 14 M5 / Air M5, MacBook Pro / Mac Studio M4 Max 16-core, Mac mini / MacBook Pro M4 Pro 14-core, Mac mini / MacBook Pro M4 Pro 12-core, MacBook Air M4 / Mac mini M4 Windows laptop Performance laptop Core Ultra 9 285H, Thin-and-light laptop Ryzen AI 9 HX 370, Gaming laptop i7-13700HX, Thin-and-light laptop R7 8845HS Desktop Desktop R7 9700X, Desktop R7 9800X3D, Desktop i5-14600KF, Desktop R7 7800X3D, Desktop R5 7500F Memory 16–32 GB ### Minimum Usable; heavy jobs up to 20% slower than the reference machine Mac MacBook Air M3, MacBook Air M2, MacBook Air M1 Windows laptop Thin-and-light laptop Core Ultra 7 258V, Gaming laptop i7-12700H, Thin-and-light laptop Core Ultra 7 155H, Thin-and-light laptop Core Ultra 5 125H, Snapdragon X Elite X1E-78-100, Thin-and-light laptop i7-1255U Desktop Desktop i5-13400F, Desktop i5-12400F, Desktop R5 5600 Memory 16 GB (8 GB for light use only) ### Not advised Heavy jobs take 25%+ longer than on the reference machine Mac — Windows laptop Thin-and-light laptop i5-1235U, Older thin-and-light laptop i5-8250U Desktop Desktop i5-10400 Memory — LADDER ## Speed ladderSpeed index: MacBook Air M4 / Mac mini M4 = 100, higher is faster, estimating total task time for the chosen scenario. Benchmark data refreshed every Monday; this run 2026-11-02 Local execution share 30% MacWindows laptopDesktopSmall print on each row: build speed / single-thread speed (reference = 100) - **MacBook Pro M5 Max 18-coreBest**2026 · Apple M5 Max 18 Core **111**faster 11% build 202 · single-thread 133 - **MacBook Pro M5 Pro 18-coreBest**2026 · Apple M5 Pro 18 Core **111**faster 11% build 201 · single-thread 132 - **MacBook Pro M5 Pro 15-coreBest**2026 · Apple M5 Pro 15 Core **110**faster 10% build 178 · single-thread 131 - **Mac Studio M3 Ultra 28-coreBest**2025 · Apple M3 Ultra 28 Core **108**faster 8% build 232 · single-thread 114 - **Desktop Core Ultra 9 285KBest**2024 · 24 cores · Intel Core Ultra 9 285K **108**faster 8% build 229 · single-thread 113 - **Desktop Core Ultra 7 265KBest**2024 · 20 cores · Intel Core Ultra 7 265K **107**faster 7% build 205 · single-thread 110 - **Desktop R9 9950XBest**2024 · 16 cores · AMD Ryzen 9 9950X **106**faster 6% build 225 · single-thread 105 - **MacBook Pro 14 M5 / Air M5Recommended**2025-26 · 10 cores · Apple M5 10 Core **106**faster 6% build 111 · single-thread 129 - **Desktop i9-14900KBest**2023 · 24 cores · Intel Core i9-14900K **106**faster 6% build 204 · single-thread 104 - **Desktop i7-14700KBest**2023 · 20 cores · Intel Core i7-14700K **104**faster 4% build 186 · single-thread 99 - **MacBook Pro / Mac Studio M4 Max 16-coreRecommended**2024-25 · Apple M4 Max 16 Core **104**faster 4% build 163 · single-thread 102 - **Desktop R7 9700XRecommended**2024 · 8 cores · AMD Ryzen 7 9700X **103**faster 3% build 142 · single-thread 103 - **Mac mini / MacBook Pro M4 Pro 14-coreRecommended**2024 · Apple M4 Pro 14 Core **103**faster 3% build 146 · single-thread 101 - **Desktop R7 9800X3DRecommended**2024 · 8 cores · AMD Ryzen 7 9800X3D **103**faster 3% build 152 · single-thread 98 - **Mac mini / MacBook Pro M4 Pro 12-coreRecommended**2024 · Apple M4 Pro 12 Core **102**faster 2% build 130 · single-thread 101 - **Performance laptop Core Ultra 9 285HRecommended**2025 · 16 cores · Intel Core Ultra 9 285H **102**faster 2% build 134 · single-thread 98 - **Desktop i5-14600KFRecommended**2023 · 14 cores · Intel Core i5-14600KF **102**faster 2% build 146 · single-thread 95 - **MacBook Air M4 / Mac mini M4Recommended**2024-25 · 10 cores · Apple M4 10 Core **100**reference build 100 · single-thread 100 - **Thin-and-light laptop Ryzen AI 9 HX 370Recommended**2024 · 12 cores · AMD Ryzen AI 9 HX 370 **99**slower 1% build 136 · single-thread 88 - **MacBook Air M3Minimum**2024 · 8 cores · Apple M3 8 Core **99**slower 1% build 85 · single-thread 105 - **Desktop R7 7800X3DRecommended**2023 · 8 cores · AMD Ryzen 7 7800X3D **98**slower 2% build 134 · single-thread 84 - **Gaming laptop i7-13700HXRecommended**2023 · 16 cores · Intel Core i7-13700HX **98**slower 2% build 126 · single-thread 84 - **Desktop R5 7500FRecommended**2023 · 6 cores · AMD Ryzen 5 7500F **97**slower 3% build 110 · single-thread 85 - **Thin-and-light laptop R7 8845HSRecommended**2024 · 8 cores · AMD Ryzen 7 8845HS **97**slower 3% build 115 · single-thread 83 - **Thin-and-light laptop Core Ultra 7 258VMinimum**2024 · 8 cores · Intel Core Ultra 7 258V **96**slower 4% build 84 · single-thread 89 - **Desktop i5-13400FMinimum**2023 · 10 cores · Intel Core i5-13400F **96**slower 4% build 104 · single-thread 81 - **Gaming laptop i7-12700HMinimum**2022 · 14 cores · Intel Core i7-12700H **95**slower 5% build 104 · single-thread 77 - **Thin-and-light laptop Core Ultra 7 155HMinimum**2024 · 16 cores · Intel Core Ultra 7 155H **94**slower 6% build 103 · single-thread 76 - **MacBook Air M2Minimum**2022 · 8 cores · Apple M2 8 Core 3500 MHz **94**slower 6% build 72 · single-thread 86 - **Desktop i5-12400FMinimum**2022 · 6 cores · Intel Core i5-12400F **93**slower 7% build 86 · single-thread 78 - **Desktop R5 5600Minimum**2022 · 6 cores · AMD Ryzen 5 5600 **92**slower 8% build 93 · single-thread 72 - **Thin-and-light laptop Core Ultra 5 125HMinimum**2024 · 14 cores · Intel Core Ultra 5 125H **92**slower 8% build 88 · single-thread 74 - **MacBook Air M1Minimum**2020 · 8 cores · Apple M1 8 Core 3200 MHz **92**slower 8% build 67 · single-thread 82 - **Snapdragon X Elite X1E-78-100Minimum**2024 · 12 cores · ARM · Qualcomm Snapdragon X Elite - X1E-78-100 **91**slower 9% build 97 · single-thread 70 - **Thin-and-light laptop i7-1255UMinimum**2022 · 10 cores · Intel Core i7-1255U **87**slower 13% build 62 · single-thread 70 - **Thin-and-light laptop i5-1235UNot advised**2022 · 10 cores · Intel Core i5-1235U **86**slower 14% build 60 · single-thread 68 - **Desktop i5-10400Not advised**2020 · 6 cores · Intel Core i5-10400 @ 2.90GHz **82**slower 18% build 59 · single-thread 57 - **Older thin-and-light laptop i5-8250UNot advised**2017-18 · 4 cores · Intel Core i5-8250U @ 1.60GHz **68**slower 32% build 33 · single-thread 42 - **Daily code modification** (local share 15%): fastest MacBook Pro M5 Max 18-core scores 105, slowest Older thin-and-light laptop i5-8250U scores 81; the slowest takes 1.3× as long for the same task. - **Bug fixes + running tests** (local share 30%): fastest MacBook Pro M5 Max 18-core scores 111, slowest Older thin-and-light laptop i5-8250U scores 68; the slowest takes 1.6× as long for the same task. - **Frontend build / large projects** (local share 45%): fastest Mac Studio M3 Ultra 28-core scores 124, slowest Older thin-and-light laptop i5-8250U scores 55; the slowest takes 2.3× as long for the same task. How to read: an index of 125 means the same task finishes about 20% faster than on the reference machine (1/1.25 of the time). “Local execution share” is the fraction of total time spent running local commands on the reference machine; model wait time is identical on every computer, so the higher the share, the bigger the gap between machines. Drag the slider to see it change. BUYING GUIDE ## Choosing a configuration **CPU: single-core sets the feel, multi-core sets build time** Tests, scripts, git and TypeScript type-checking mostly use one core; npm installs, Next.js/Vite production builds and parallel tests use many. If you mostly edit code, look at single-core; if you build large projects often, look at multi-core. **Memory: 4 GB is the official floor, 16 GB is our advice** Codex asks for 4 GB (8 GB recommended), but that covers the tool only. An agent also runs dev servers, a Playwright browser and test processes, which strain 8 GB machines (this one is our recommendation, not an official figure). **Storage: SSD is the floor, small-file speed matters** One npm install writes tens of thousands of small files. A hard disk or low-end eMMC makes “installing dependencies” the slowest step. **Network: the rest of the time is waiting for the model** A faster computer only compresses the local part. Waiting time depends on model speed and your network path to the API; an unstable network amplifies it. FAQ ## FAQ **What computer do I need to run Codex or Claude Code?** Official minimums: Codex needs macOS 12+, Ubuntu 20.04+/Debian 10+ or Windows 11 with WSL2, and 4 GB RAM (8 GB recommended). Claude Code needs macOS 13+, Windows 10 1809+, Ubuntu 20.04+, Debian 10+ or Alpine 3.19+, 4 GB+ RAM and an x64 or ARM64 CPU. Those are only the tools’ own limits. An agent also runs installs, builds, tests and a browser, so we suggest 16 GB RAM and a CPU at least in this page’s “Minimum” tier; pick “Recommended” for a smooth experience. **Will a MacBook Air run Codex or Claude Code?** Yes. By this page’s tiers, MacBook Air M3, MacBook Air M2, MacBook Air M1 sit in “Minimum” and MacBook Pro 14 M5 / Air M5, MacBook Pro / Mac Studio M4 Max 16-core, Mac mini / MacBook Pro M4 Pro 14-core, Mac mini / MacBook Pro M4 Pro 12-core, MacBook Air M4 / Mac mini M4 in “Recommended”. 8 GB models are fine for light use but run short of memory when a dev server and browser tests run together. **How much faster is a faster computer for AI coding?** It depends on how much of the task runs locally. Between the fastest and slowest mainstream machines, the same task differs by: Daily code modification: up to 1.3×; Bug fixes + running tests: up to 1.6×; Frontend build / large projects: up to 2.3×. Model inference runs in the cloud and does not change with your computer. **Why doesn’t a top-end machine help more?** Most of the time is spent waiting for the model. Third-party papers measure local tool execution at roughly 2–39% of an agent’s end-to-end time, the rest being model inference and network. Moving from “Recommended” to “Best” speeds everyday tasks by only ~5–10%; build-heavy and test-heavy projects benefit more. **More questions (4)Anything to watch on Windows?** Codex recommends Windows 11 with WSL2. Keep projects inside the Linux filesystem (/home/…), not under /mnt/c; Microsoft documents a large slowdown across the boundary. For native Windows development, put the project and package caches on a Dev Drive. **How accurate is the estimate?** Mac: trustworthy. Checked against real compile times from the community XcodeBenchmark (20 Apple chips on Xcode 16): median error 6.8%, max 18.8%; on Xcode 26 data (including M5) median 3.5%. Windows laptops and desktops: lower confidence, since no public compile-time dataset exists to run the same check; real results are rarely better than the estimate. **How is the speed index calculated?** Speed index = 100 ÷ [(1 − local share) + local share × local-time factor]. The local-time factor weights “build” and “single-thread” parts per scenario: build speed = (multi-core score ÷ reference multi-core score)^0.79 (exponent fitted to XcodeBenchmark), single-thread speed = single-thread score ÷ reference single-thread score. **What are the assumptions and limits?** The three scenarios’ local shares (15% / 30% / 45%) and build weights (20% / 30% / 70%) are our assumptions based on ranges reported in papers, not measurements. Memory pressure, thermal throttling, battery mode, disk and OS differences are not modeled. This is a buying reference, not measured Codex results. Sources**benchmarks and official docs** - [PassMark CPU Benchmarks (single-thread / multi-thread scores)](https://www.cpubenchmark.net/singleThread.html) fetched 2026-11-02 - [XcodeBenchmark (community-submitted real Mac compilation time)](https://github.com/devMEremenko/XcodeBenchmark) fetched 2026-11-02 - [OpenAI Codex official installation documentation (system requirements)](https://github.com/openai/codex/blob/main/docs/install.md) fetched 2026-09-24 - [Microsoft Learn: WSL cross-filesystem performance](https://learn.microsoft.com/en-us/windows/wsl/filesystems) fetched 2026-09-24 - [Microsoft Learn: Dev Drive](https://learn.microsoft.com/en-us/windows/dev-drive/) fetched 2026-09-24 - [Claude Code official documentation (system requirements)](https://code.claude.com/docs/en/setup) fetched 2026-09-24