CPU Performance Rankings

3,400+ processors ranked by averaged PassMark, Geekbench, Cinebench and 3DMark results

The most complete processor ranking you can check in one place

Choosing a processor in 2026 is harder than the marketing makes it look. Every launch cycle brings new names, new core counts, and new claims about artificial intelligence performance and gaming leadership, and the difference between a good purchase and an expensive mistake usually comes down to numbers the box never mentions. This ranking exists to make that decision measurable. Every processor below is positioned by averaged benchmark results from the major industry test suites, including PassMark, Geekbench, Cinebench, and 3DMark's physics tests, consolidated into a single comparable score. The database tracks more than 3,400 processors released over the past two decades, from current flagship silicon to the legacy chips still running in millions of working PCs, and the same scoring method is applied to all of them without favoritism toward either manufacturer.

What makes this leaderboard different from a typical benchmark chart is that the numbers connect to real gaming outcomes. Behind this page sits a measured performance library covering 53 tested games, more than 190,000 unique CPU and graphics card combinations, and over two million recorded frame rate results. When a ranking says a processor is fast, that claim can be checked against how the same chip actually performed in games at 1080p, 1440p, and 4K across Low, Medium, High, and Ultra settings. Benchmarks rank the silicon. Measured frame rates show what the silicon does for you.

How the ranking score is calculated

Each processor's score is the average of its normalized results across every benchmark suite in the database, weighted by how commonly that suite is cited in reviews and purchase decisions. Averaging matters, because any single test can flatter a specific design. A chip with very wide cores and modest clock speeds scores brilliantly in Cinebench's multi-core rendering runs and looks ordinary in tests that reward per-core throughput. A high-clocked gaming-focused design does the reverse. Averaging across suites produces a balanced picture of overall capability, which is the right default for a general ranking, and the per-processor detail pages break the scores back out by suite for anyone who cares about a specific workload.

Two honest limitations are worth stating up front. First, laptop processors with configurable power limits can score differently depending on how a manufacturer tunes a specific machine, so mobile results represent the typical configuration rather than the best case. Second, average scores say nothing about price. A processor that ranks 20 places higher but costs three times as much is not a better purchase for most people, which is why a dedicated price-performance leaderboard exists alongside this page and why several sections below talk about value explicitly.

What the leaderboards below show

The page is organized the way people actually shop. The overall table at the top ranks every tracked processor by consolidated score, with Intel and AMD filters for brand comparisons. Below it, the desktop, laptop and mobile, and workstation and server sections each rank the top 60 chips in that class, because a laptop processor competing against an unlocked desktop flagship on watts-adjusted terms is a comparison nobody shopping for a thin-and-light ever asked for. Every section pairs its ranking table with a visual chart of the top ten, a search box that filters the list as you type, and a load more button that extends the table without leaving the page. There are no page two URLs here; the entire ranking lives on one page so search engines and readers see the same complete picture.

#1 Overall MSRP $14,813

AMD EPYC 9965

AMD • Released 2024

CORES 192 (384T)
CLOCK 2.3 GHz
TDP 500W
Benchmark Score 620,487
View Details
#2 Overall MSRP $13,564

AMD EPYC 9845

AMD • Released 2024

CORES 160 (320T)
CLOCK 2.1 GHz
TDP 390W
Benchmark Score 523,613
View Details
#3 Overall MSRP $12,984

AMD EPYC 9755

AMD • Released 2024

CORES 128 (256T)
CLOCK 2.7 GHz
TDP 500W
Benchmark Score 505,778
View Details
#4 Overall MSRP $12,141

AMD EPYC 9745

AMD • Released 2024

CORES 128 (256T)
CLOCK 2.4 GHz
TDP 400W
Benchmark Score 425,973
View Details

Desktop CPU Performance Hierarchy Desktop

7 83,807
8 81,127
9 79,371
10 79,097
11 75,779
12 75,488
13 74,640
14 71,910
15 70,879
16 70,163
17 69,996
18 69,515
19 69,355
20 66,855
22 66,099
23 65,914
24 64,640
25 64,438

Laptop CPU Performance Hierarchy Laptop

1 106,413
2 97,453
3 91,199
5 81,103
9 76,212
10 76,155
12 71,797
13 69,875
14 67,469
16 63,173
18 62,738
19 62,498
21 60,104
22 60,099
23 59,641
24 58,826

The overall leaderboard

Every tracked processor, every class, one consolidated score. Desktop flagships, mobile efficiency leaders, and throughput-first workstation parts compete directly here, which is exactly why the category sections below also exist: this table answers which silicon is fastest overall, and the sections answer which is best for the machine you are actually building. Use the Intel and AMD tabs above to narrow the field.

Overall CPU Leaderboard

RANK CPU MODEL CORES CLOCK BENCHMARK SCORE TDP (W) ACTION
#1 192C/384T — 620,487 500W DETAILS
#2 160C/320T — 523,613 390W DETAILS
#3 128C/256T — 505,778 500W DETAILS
#4 128C/256T — 425,973 400W DETAILS
#5 96C/192T — 412,068 350W DETAILS
#6 96C/192T — 397,773 400W DETAILS
#7 64C/128T — 379,408 300W DETAILS
#8 96C/192T — 373,479 400W DETAILS
#9 72C/144T — 365,194 500W DETAILS
#10 128C/256T — 364,371 360W DETAILS
#11 48C/96T — 350,933 400W DETAILS
#12 64C/128T — 321,753 350W DETAILS
#13 64C/128T — 320,749 350W DETAILS
#14 80C/160T — 315,524 350W DETAILS
#15 64C/128T — 311,774 400W DETAILS
#16 112C/224T — 310,619 340W DETAILS
#17 64C/128T — 300,372 350W DETAILS
#18 64C/128T — 287,066 360W DETAILS
#19 64C/128T — 286,102 350W DETAILS
#20 72C/144T — 285,471 400W DETAILS
#21 144C/144T — 280,438 330W DETAILS
#22 32C/64T — 279,778 350W DETAILS
#23 96C/192T — 266,914 400W DETAILS
#24 128C/256T — 251,516 500W DETAILS
#25 84C/168T — 244,274 290W DETAILS
#26 48C/96T — 238,263 330W DETAILS
#27 48C/96T — 217,854 300W DETAILS
#28 60C/120T — 209,881 385W DETAILS
#29 48C/96T — 194,901 300W DETAILS
#30 32C/64T — 194,228 210W DETAILS
#31 48C/96T — 193,477 300W DETAILS
#32 96C/96T — 187,718 250W DETAILS
#33 64C/128T — 185,092 200W DETAILS
#34 32C/64T — 182,700 350W DETAILS
#35 64C/128T — 179,916 280W DETAILS
#36 48C/96T — 176,227 270W DETAILS
#37 64C/128T — 171,748 280W DETAILS
#38 64C/128T — 167,788 225W DETAILS
#39 32C/64T — 162,497 320W DETAILS
#40 56C/112T — 161,973 240W DETAILS
#41 32C/64T — 160,358 280W DETAILS
#42 32C/64T — 158,540 275W DETAILS
#43 32C/64T — 154,858 300W DETAILS
#44 24C/48T — 147,009 350W DETAILS
#45 48C/96T — 146,881 200W DETAILS
#46 16C/32T — 145,728 170W DETAILS
#47 48C/96T — 144,824 225W DETAILS
#48 44C/88T — 144,323 340W DETAILS
#49 16C/32T — 143,735 170W DETAILS
#50 32C/64T — 143,444 350W DETAILS

The most complete processor ranking you can check in one place

Choosing a processor in 2026 is harder than the marketing makes it look. Every launch cycle brings new names, new core counts, and new claims about artificial intelligence performance and gaming leadership, and the difference between a good purchase and an expensive mistake usually comes down to numbers the box never mentions. This ranking exists to make that decision measurable. Every processor below is positioned by averaged benchmark results from the major industry test suites, including PassMark, Geekbench, Cinebench, and 3DMark's physics tests, consolidated into a single comparable score. The database tracks more than 3,400 processors released over the past two decades, from current flagship silicon to the legacy chips still running in millions of working PCs, and the same scoring method is applied to all of them without favoritism toward either manufacturer.

What makes this leaderboard different from a typical benchmark chart is that the numbers connect to real gaming outcomes. Behind this page sits a measured performance library covering 53 tested games, more than 190,000 unique CPU and graphics card combinations, and over two million recorded frame rate results. When a ranking says a processor is fast, that claim can be checked against how the same chip actually performed in games at 1080p, 1440p, and 4K across Low, Medium, High, and Ultra settings. Benchmarks rank the silicon. Measured frame rates show what the silicon does for you.

How the ranking score is calculated

Each processor's score is the average of its normalized results across every benchmark suite in the database, weighted by how commonly that suite is cited in reviews and purchase decisions. Averaging matters, because any single test can flatter a specific design. A chip with very wide cores and modest clock speeds scores brilliantly in Cinebench's multi-core rendering runs and looks ordinary in tests that reward per-core throughput. A high-clocked gaming-focused design does the reverse. Averaging across suites produces a balanced picture of overall capability, which is the right default for a general ranking, and the per-processor detail pages break the scores back out by suite for anyone who cares about a specific workload.

Two honest limitations are worth stating up front. First, laptop processors with configurable power limits can score differently depending on how a manufacturer tunes a specific machine, so mobile results represent the typical configuration rather than the best case. Second, average scores say nothing about price. A processor that ranks 20 places higher but costs three times as much is not a better purchase for most people, which is why a dedicated price-performance leaderboard exists alongside this page and why several sections below talk about value explicitly.

What the leaderboards below show

The page is organized the way people actually shop. The overall table at the top ranks every tracked processor by consolidated score, with Intel and AMD filters for brand comparisons. Below it, the desktop, laptop and mobile, and workstation and server sections each rank the top 60 chips in that class, because a laptop processor competing against an unlocked desktop flagship on watts-adjusted terms is a comparison nobody shopping for a thin-and-light ever asked for. Every section pairs its ranking table with a visual chart of the top ten, a search box that filters the list as you type, and a load more button that extends the table without leaving the page. There are no page two URLs here; the entire ranking lives on one page so search engines and readers see the same complete picture.

Desktop processors: where the performance ceiling lives

The desktop section is where raw capability is decided. Unlocked by power budgets that mobile designs can only dream about, the current generation of desktop flagships has pushed core counts and cache sizes into territory that professional workstations occupied five years ago, at prices mainstream builders can reach. The top of this table is dominated by the newest flagship parts, and the score gaps between them are real but small; the far more interesting action happens ten to forty places down, where previous-generation flagships now sell at prices that make them some of the strongest value plays in the table.

For gaming specifically, the received wisdom that cores do not matter beyond a certain point remains mostly true, with one modern correction: the newest game engines with heavy simulation and crowd systems will visibly use everything a modern 8-core processor offers, and the measured game data shows steeper minimum frame rates on the high-core-count parts in exactly those titles. The practical guidance is unchanged, though. For a pure gaming build, buy the processor that leads in single-thread throughput and put the savings into the graphics card, because at the resolutions most people actually play, the graphics card remains the decisive component in the large majority of the 53 games in the test library.

Desktop Processor Leaderboard

RANK CPU MODEL CORES CLOCK BENCHMARK SCORE TDP (W) ACTION
#1 64C/128T 0.1 GHz 321,753 350W DETAILS
#2 32C/64T 0.1 GHz 279,778 350W DETAILS
#3 64C/128T 0.1 GHz 171,748 280W DETAILS
#4 16C/32T 0.1 GHz 101,041 350W DETAILS
#5 24C/24T 0.1 GHz 93,785 125W DETAILS
#6 24C/24T 0.1 GHz 84,003 125W DETAILS
#7 24C/24T 0.1 GHz 83,807 125W DETAILS
#8 24C/32T 0.1 GHz 81,127 150W DETAILS
#9 24C/32T 0.1 GHz 79,371 125W DETAILS
#10 24C/32T 0.1 GHz 79,097 125W DETAILS
#11 16C/32T 0.1 GHz 75,779 170W DETAILS
#12 24C/24T 0.1 GHz 75,488 65W DETAILS
#13 16C/32T 0.1 GHz 74,640 170W DETAILS
#14 20C/20T 0.1 GHz 71,910 125W DETAILS
#15 20C/20T 0.1 GHz 70,879 125W DETAILS
#16 20C/28T 0.1 GHz 70,163 125W DETAILS
#17 8C/16T 0.1 GHz 69,996 65W DETAILS
#18 16C/32T 0.1 GHz 69,515 170W DETAILS
#19 20C/28T 0.1 GHz 69,355 125W DETAILS
#20 18C/18T 0.1 GHz 66,855 125W DETAILS
#21 18C/18T 0.1 GHz 66,159 125W DETAILS
#22 12C/24T 0.1 GHz 66,099 125W DETAILS
#23 16C/32T 0.1 GHz 65,914 120W DETAILS
#24 20C/20T 0.1 GHz 64,640 65W DETAILS
#25 20C/20T 0.1 GHz 64,438 65W DETAILS
#26 24C/32T 0.1 GHz 64,051 150W DETAILS
#27 8C/16T 0.1 GHz 63,331 65W DETAILS
#28 16C/24T 0.1 GHz 63,080 65W DETAILS
#29 24C/32T 0.1 GHz 61,841 125W DETAILS
#30 24C/32T 0.1 GHz 61,766 125W DETAILS
#31 24C/32T 0.1 GHz 61,723 35W DETAILS
#32 24C/32T 0.1 GHz 60,676 65W DETAILS
#33 24C/32T 0.1 GHz 60,008 65W DETAILS
#34 8C/16T 0.1 GHz 58,386 120W DETAILS
#35 24C/32T 0.1 GHz 58,115 65W DETAILS
#36 12C/24T 0.1 GHz 57,498 120W DETAILS
#37 16C/32T 0.1 GHz 56,555 280W DETAILS
#38 10C/20T 0.1 GHz 55,919 125W DETAILS
#39 14C/14T 0.1 GHz 55,093 125W DETAILS
#40 12C/24T 0.1 GHz 54,762 120W DETAILS
#41 14C/14T 0.1 GHz 54,053 125W DETAILS
#42 20C/28T 0.1 GHz 53,620 65W DETAILS
#43 12C/24T 0.1 GHz 53,288 170W DETAILS
#44 6C/12T 0.1 GHz 52,873 65W DETAILS
#45 20C/28T 0.1 GHz 52,301 65W DETAILS
#46 16C/32T 0.1 GHz 51,947 105W DETAILS
#47 24C/32T 0.1 GHz 51,730 65W DETAILS
#48 24C/24T 0.1 GHz 51,310 35W DETAILS
#49 24C/32T 0.1 GHz 51,015 35W DETAILS
#50 16C/32T 0.1 GHz 50,718 105W DETAILS
#51 12C/24T 0.1 GHz 49,845 65W DETAILS
#52 14C/20T 0.1 GHz 49,394 125W DETAILS
#53 12C/24T 0.1 GHz 49,228 65W DETAILS
#54 8C/16T 0.1 GHz 49,196 65W DETAILS
#55 14C/14T 0.1 GHz 48,995 65W DETAILS
#56 14C/20T 0.1 GHz 48,618 125W DETAILS
#57 14C/14T 0.1 GHz 48,201 65W DETAILS
#58 12C/24T 0.1 GHz 47,918 65W DETAILS
#59 12C/24T 0.1 GHz 47,908 120W DETAILS
#60 20C/20T 0.1 GHz 47,697 35W DETAILS

Laptop and mobile processors: efficiency as a feature

Mobile silicon plays a different sport. A laptop processor has to deliver its performance inside a thermal envelope that a desktop chip would burn through in seconds, which is why the mobile leaderboard rewards designs that sustain high clocks efficiently rather than designs that simply post a big score in a short burst. The current generation of mobile processors is genuinely impressive on this measure: thin-and-light machines now run competitive esports titles at high refresh rates on integrated graphics alone, something that required a discrete card not long ago, and the newer processors with dedicated neural hardware handle background artificial intelligence features, video calls, and content creation work with headroom to spare.

When comparing mobile chips, pay attention to the configured power, not just the model name. The same processor family ships in configurations ranging from about 15 watts in ultrabooks to 55 watts and beyond in gaming and workstation laptops, and sustained performance scales accordingly. The scores in the mobile section reflect typical configurations; a gaming laptop running the same chip at its maximum power setting will outperform it, and a fanless tablet will land below it. For buyers deciding between two similarly named chips, the rank order here is a more reliable guide than the marketing suffixes, which have never told the whole story.

Laptop & Mobile Processor Leaderboard

RANK CPU MODEL CORES CLOCK BENCHMARK SCORE TDP (W) ACTION
#1 12C/24T 0.1 GHz 106,413 55W DETAILS
#2 16C/32T 0.1 GHz 97,453 55W DETAILS
#3 16C/32T 0.1 GHz 91,199 55W DETAILS
#4 12C/24T 0.1 GHz 90,541 55W DETAILS
#5 16C/32T 0.1 GHz 81,103 55W DETAILS
#6 16C/32T 0.1 GHz 80,762 55W DETAILS
#7 24C/24T 0.1 GHz 79,574 55W DETAILS
#8 16C/32T 0.1 GHz 77,740 55W DETAILS
#9 16C/32T 0.1 GHz 76,212 55W DETAILS
#10 24C/24T 0.1 GHz 76,155 55W DETAILS
#11 12C/24T 0.1 GHz 74,496 28W DETAILS
#12 12C/24T 0.1 GHz 71,797 55W DETAILS
#13 16C/32T 0.1 GHz 69,875 55W DETAILS
#14 24C/24T 0.1 GHz 67,469 55W DETAILS
#15 12C/24T 0.1 GHz 63,765 55W DETAILS
#16 20C/20T 0.1 GHz 63,173 55W DETAILS
#17 12C/24T 0.1 GHz 62,839 28W DETAILS
#18 20C/20T 0.1 GHz 62,738 55W DETAILS
#19 10C/20T 0.1 GHz 62,498 28W DETAILS
#20 20C/20T 0.1 GHz 61,834 55W DETAILS
#21 8C/16T 0.1 GHz 60,104 55W DETAILS
#22 16C/32T 0.1 GHz 60,099 55W DETAILS
#23 16C/32T 0.1 GHz 59,641 55W DETAILS
#24 12C/24T 0.1 GHz 58,826 28W DETAILS
#25 12C/24T 0.1 GHz 56,306 28W DETAILS
#26 12C/24T 0.1 GHz 56,273 55W DETAILS
#27 24C/32T 0.1 GHz 56,004 55W DETAILS
#28 8C/16T 0.1 GHz 52,901 28W DETAILS
#29 14C/14T 0.1 GHz 52,073 55W DETAILS
#30 20C/28T 0.1 GHz 50,761 55W DETAILS
#31 12C/24T 0.1 GHz 50,448 28W DETAILS
#32 24C/32T 0.1 GHz 50,398 55W DETAILS
#33 8C/16T 0.1 GHz 49,796 55W DETAILS
#34 14C/14T 0.1 GHz 48,287 55W DETAILS
#35 6C/12T 0.1 GHz 48,171 55W DETAILS
#36 16C/16T 0.1 GHz 47,468 25W DETAILS
#37 12C/24T 0.1 GHz 47,022 28W DETAILS
#38 12C/24T 0.1 GHz 46,654 55W DETAILS
#39 24C/32T 0.1 GHz 46,098 55W DETAILS
#40 12C/24T 0.1 GHz 46,030 28W DETAILS
#41 20C/28T 0.1 GHz 45,953 55W DETAILS
#42 16C/16T 0.1 GHz 44,466 25W DETAILS
#43 24C/32T 0.1 GHz 44,342 55W DETAILS
#44 8C/16T 0.1 GHz 44,309 55W DETAILS
#45 8C/16T 0.1 GHz 43,717 45W DETAILS
#46 8C/16T 0.1 GHz 43,689 35W DETAILS
#47 10C/20T 0.1 GHz 43,431 28W DETAILS
#48 8C/16T 0.1 GHz 43,326 55W DETAILS
#49 16C/16T 0.1 GHz 43,210 25W DETAILS
#50 16C/24T 0.1 GHz 42,487 55W DETAILS
#51 8C/16T 0.1 GHz 41,963 45W DETAILS
#52 16C/24T 0.1 GHz 41,779 55W DETAILS
#53 16C/16T 0.1 GHz 41,621 28W DETAILS
#54 16C/24T 0.1 GHz 41,576 55W DETAILS
#55 16C/16T 0.1 GHz 41,263 25W DETAILS
#56 16C/16T 0.1 GHz 41,215 25W DETAILS
#57 6C/12T 0.1 GHz 41,208 28W DETAILS
#58 16C/16T 0.1 GHz 40,967 25W DETAILS
#59 16C/16T 0.1 GHz 40,518 25W DETAILS
#60 8C/16T 0.1 GHz 40,431 35W DETAILS

Workstation and server processors: throughput first

The workstation and server section ranks the parts built for sustained multi-threaded throughput: render nodes, compile farms, virtualization hosts, scientific computation, and increasingly, local artificial intelligence inference. These chips trade clock speed and gaming responsiveness for core counts that start where desktop parts stop, and their benchmark scores reflect that choice. In multi-core rendering and computation suites they dominate the overall leaderboard, and in single-thread tests they sit well below gaming desktops, which is exactly the tradeoff their buyers want.

One trend in this section deserves attention from a wider audience than its traditional buyer. The used and previous-generation workstation market has become a genuinely practical route to a heavily multi-threaded machine at mainstream prices, and because every chip in this table is ranked by the same method, a shopper can directly compare a previous-generation server part against a current desktop flagship and see precisely how much multi-core throughput the older chip still delivers per dollar. For rendering, code compilation, and video encoding workloads, that comparison regularly surprises people who assumed newer always means faster for their workload.

Workstation & Server Leaderboard

RANK CPU MODEL CORES CLOCK BENCHMARK SCORE TDP (W) ACTION
#1 192C/384T 0.1 GHz 620,487 500W DETAILS
#2 160C/320T 0.1 GHz 523,613 390W DETAILS
#3 128C/256T 0.1 GHz 505,778 500W DETAILS
#4 128C/256T 0.1 GHz 425,973 400W DETAILS
#5 96C/192T 0.1 GHz 412,068 350W DETAILS
#6 96C/192T 0.1 GHz 397,773 400W DETAILS
#7 64C/128T 0.1 GHz 379,408 300W DETAILS
#8 96C/192T 0.1 GHz 373,479 400W DETAILS
#9 72C/144T 0.1 GHz 365,194 500W DETAILS
#10 128C/256T 0.1 GHz 364,371 360W DETAILS
#11 48C/96T 0.1 GHz 350,933 400W DETAILS
#12 64C/128T 0.1 GHz 320,749 350W DETAILS
#13 80C/160T 0.1 GHz 315,524 350W DETAILS
#14 64C/128T 0.1 GHz 311,774 400W DETAILS
#15 112C/224T 0.1 GHz 310,619 340W DETAILS
#16 64C/128T 0.1 GHz 300,372 350W DETAILS
#17 64C/128T 0.1 GHz 287,066 360W DETAILS
#18 64C/128T 0.1 GHz 286,102 350W DETAILS
#19 72C/144T 0.1 GHz 285,471 400W DETAILS
#20 144C/144T 0.1 GHz 280,438 330W DETAILS
#21 96C/192T 0.1 GHz 266,914 400W DETAILS
#22 128C/256T 0.1 GHz 251,516 500W DETAILS
#23 84C/168T 0.1 GHz 244,274 290W DETAILS
#24 48C/96T 0.1 GHz 238,263 330W DETAILS
#25 48C/96T 0.1 GHz 217,854 300W DETAILS
#26 60C/120T 0.1 GHz 209,881 385W DETAILS
#27 48C/96T 0.1 GHz 194,901 300W DETAILS
#28 32C/64T 0.1 GHz 194,228 210W DETAILS
#29 48C/96T 0.1 GHz 193,477 300W DETAILS
#30 96C/96T 0.1 GHz 187,718 250W DETAILS
#31 64C/128T 0.1 GHz 185,092 200W DETAILS
#32 32C/64T 0.1 GHz 182,700 350W DETAILS
#33 64C/128T 0.1 GHz 179,916 280W DETAILS
#34 48C/96T 0.1 GHz 176,227 270W DETAILS
#35 64C/128T 0.1 GHz 167,788 225W DETAILS
#36 32C/64T 0.1 GHz 162,497 320W DETAILS
#37 56C/112T 0.1 GHz 161,973 240W DETAILS
#38 32C/64T 0.1 GHz 160,358 280W DETAILS
#39 32C/64T 0.1 GHz 158,540 275W DETAILS
#40 32C/64T 0.1 GHz 154,858 300W DETAILS
#41 24C/48T 0.1 GHz 147,009 350W DETAILS
#42 48C/96T 0.1 GHz 146,881 200W DETAILS
#43 16C/32T 0.1 GHz 145,728 170W DETAILS
#44 48C/96T 0.1 GHz 144,824 225W DETAILS
#45 44C/88T 0.1 GHz 144,323 340W DETAILS
#46 16C/32T 0.1 GHz 143,735 170W DETAILS
#47 32C/64T 0.1 GHz 143,444 350W DETAILS
#48 28C/56T 0.1 GHz 143,103 270W DETAILS
#49 32C/64T 0.1 GHz 140,694 270W DETAILS
#50 24C/48T 0.1 GHz 133,174 320W DETAILS
#51 64C/64T 0.1 GHz 129,930 205W DETAILS
#52 32C/64T 0.1 GHz 126,810 280W DETAILS
#53 32C/64T 0.1 GHz 124,756 250W DETAILS
#54 32C/64T 0.1 GHz 124,171 280W DETAILS
#55 48C/96T 0.1 GHz 124,006 225W DETAILS
#56 32C/64T 0.1 GHz 120,427 320W DETAILS
#57 32C/64T 0.1 GHz 118,307 335W DETAILS
#58 24C/48T 0.1 GHz 116,388 200W DETAILS
#59 24C/48T 0.1 GHz 116,060 250W DETAILS
#60 24C/48T 0.1 GHz 115,190 255W DETAILS

Reading the table for your workload

Rankings answer which is fastest. They do not answer which is right for you, and the honest answer depends on the work. For high refresh rate competitive gaming, single-thread score and cache matter most, and any processor in the overall top forty will drive a strong graphics card without meaningful bottlenecking at 1440p and above. For streaming while gaming, encoding video, or running heavy background tasks alongside play, core count rises in importance and the eight-core-and-up region of the desktop table earns its premium. For content creation, the multi-core leaders in the workstation section set the pace for rendering and export times. For general productivity and office work, nearly anything in the upper half of any table is more processor than email, documents, and browsing will ever ask for, and value should drive the decision instead of rank.

Gaming performance and the processor rankings

Because the ranking is grounded in a measured game library, a note on how processor choice actually affects frame rates is worth the space. At 4K, the graphics card does nearly all the work, and processor rank barely moves the result; measured differences between top-tier processors at 4K Ultra in the game library typically sit within a few percent. At 1440p the gap widens slightly. At 1080p, especially at High and Ultra settings where the graphics card is not the limiting factor, faster processors pull visibly ahead, and that is the resolution where the top of this table pays for itself on a high refresh rate monitor. Anyone pairing a premium processor with a modest graphics card at 1080p should expect the graphics card to be the ceiling, and the upgrade planning tools elsewhere on the site will show exactly that tradeoff for any specific pairing.

Cooling, power, and the real cost of a top rank

The benchmark score is the price of admission, not the whole bill. High-ranking desktop flagships increasingly assume serious cooling to sustain their boost clocks, and the difference between a chip's headline score and its all-day sustained performance is exactly the difference a good cooler recovers. The TDP column in each table is the honest place to start: it approximates the sustained heat the processor asks a cooling solution to absorb, and shopping for a cooler rated comfortably above that number is how a top-ranked chip keeps delivering top-ranked results through long sessions rather than in short bursts. Power supply headroom follows the same logic, particularly for builds pairing a flagship processor with a flagship graphics card, where transient spikes from both components at once can trip a supply sized to the averages alone.

Platform cost is the quiet budget killer. A new-generation flagship typically requires a current chipset board and current memory, and moving the same money one or two tiers down the ranking frequently funds a full platform, or a meaningfully faster graphics card, with a processor that still ranks inside the top ten percent of everything ever tested here. The leaderboard makes that tradeoff visible in a way marketing pages never do, because the rank number tells you exactly how much performance the extra money is buying, and how much of it is going to a name.

Generational leaps: what a new generation actually buys

Looking down the overall table across generations produces a consistent pattern worth internalizing. A new generation typically buys one clear step of improvement at the same price tier, better efficiency at the same performance, or more cache where the architecture can exploit it. Two generations back, the flagship of that era usually sits alongside the current generation's midrange, at prices that have fallen accordingly. Three generations back is where used-market value concentrates, especially for core-heavy parts whose multi-threaded throughput has aged far better than their gaming rank suggests. None of these patterns is a rule; every part is ranked on its own measured results right here, which is precisely why a data-backed leaderboard ages better than advice built on launch-day impressions.

For upgrade planners the table doubles as a sanity check. Find your current processor's rank, find the rank of the part you are considering, and the gap between the two is the honest ceiling of what the upgrade can deliver, before the graphics card, memory, or storage have their say. If that gap is small at the resolutions you play, the measured game data on this site will usually confirm that the money is better spent elsewhere in the build.

The measurement backbone behind this ranking

Numbers are only as good as their collection method, so here is exactly what sits behind this page. The benchmark database aggregates results from the industry's standard test suites across every processor tracked, normalized so that suites with different scales contribute comparably. On top of that, the game library holds measured frame rates from 53 titles, recorded across more than 190,000 distinct processor and graphics card pairings at three resolutions and four settings presets, over two million individual results in total. The consequence is practical: when this leaderboard says a processor ranks highly, that claim is cross-checkable against how the same silicon performed in real game runs on real machines, not just in synthetic loops. Where a synthetic score and the measured game data ever disagree sharply, the game data is what should win your trust, and the per-game pages exist so you can perform exactly that check for any combination you care about.

Using this page with the rest of the tools

The leaderboard answers half of a purchase decision; the interactive tools on this site answer the other half. After narrowing your processor shortlist here, the comparison tool pits any two chips head to head on their shared benchmark suites. The Can I Run It tool tests any processor and graphics card pairing against a specific game using measured results where they exist. Rate My PC scores a complete build and identifies which component limits it. And the price-performance leaderboard converts every score on this page into performance per dollar at launch MSRP. A workflow that starts at this ranking and ends at a measured pairing check is the shortest honest path from what is fastest to what is right for the machine you are about to build or buy.

Frequently asked questions

How often is the CPU leaderboard updated? The underlying benchmark database is updated continuously as new results and new processors are added, and this page's rankings are rebuilt from that data on a short cache cycle, so newly launched chips appear quickly and scores for existing parts adjust as additional test data arrives.

Why does a more expensive processor sometimes rank below a cheaper one? The ranking is purely performance-based. Price-performance is a separate leaderboard, and the two pages disagree by design, because value depends on both sides of that fraction.

Are Intel and AMD processors scored the same way? Yes. Both manufacturers' parts go through identical averaging across the same benchmark suites, and the brand filter above exists purely for convenience, not for separate scoring.

Do these rankings predict game performance directly? They predict the processor's contribution. Actual game frame rates also depend on the graphics card, resolution, and settings, and the per-game pages on this site measure those combinations directly across more than 190,000 tested hardware pairings.

What about overclocking? Rankings reflect stock performance. Overclocking headroom varies by individual chip and cooling, so treat it as potential upside on unlocked parts rather than something a general leaderboard can promise.