Hardware · Clusters · Example cluster

Four Raspberry Pi 5s (16 GB): what it runs together

Four 16 GB Raspberry Pi 5 boards on a gigabit switch.

A cheap way to learn how clustering works. Distributed Llama splits every layer across all four boards, so small models type more than twice as fast as on one Pi (about 14 tokens per second for Llama 3.2 3B). Pipeline tools only add memory: big models fit, but at well under a token per second.

Change the devices How clustering works

54.4 GBusable memory together (the largest device has 13.6 GB)

107models only the cluster can run

133run better or faster than on any one device

118B (6.0B active)biggest model it adds: Groot 65B A6B 261023

Connected by 1 Gb Ethernet. The wait for the first word is worked out for a 1,000-token prompt (a short chat).

The devices

Tool to use

llama.cpp RPC best estimate for 182 of the 240 models the cluster adds

Mixed devices you already own, a Mac, a gaming PC, a mini PC or a Jetson, when one GGUF model won't fit on any of them alone.

ggml-rpc-server --host 0.0.0.0 --port 50052 -c   # on each worker
llama-server -hf <model-GGUF> -ngl 99 --rpc 192.168.1.10:50052,192.168.1.11:50052   # on the main machine

Also works. The number is how many of the added models each one runs:

467 LLMs this cluster can run

Models the cluster adds come first, best match first, then the ones a single device already runs as well. Filter vs one device to pick. Speeds are estimates for one person chatting.

Quality Model Quant Memory Speed Runs Devices Tool vs one device
88Excellent quality (A+) · 88 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. Qwen3.6 35B-A3BAlibaba · 36B (3.0B active) Q4_K_M 28.8 GB 15.2 tok/s Runs well Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Tensor split Distributed Llama Needs the clusterWon’t run on any one of them
88Excellent quality (A+) · 88 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. Qwen3.5 35B-A3BAlibaba · 36B (3.0B active) Q4_K_M 28.8 GB 15.2 tok/s Runs well Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Tensor split Distributed Llama Needs the clusterWon’t run on any one of them
92Excellent quality (A+) · 92 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 86/100 (quality, fit, speed and memory use) sets the ranking. MiniCPM-V 4.7 35B-A3BOpenBMB · 35B (3.0B active) Q6_K 34.1 GB 3.7 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
91Excellent quality (A+) · 91 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking. Nex N2 Mini (35B-A3B)Nex-AGI · 35B (3.0B active) Q6_K 34.0 GB 3.7 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
90Excellent quality (A+) · 90 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 85/100 (quality, fit, speed and memory use) sets the ranking. KAT-Coder V2.5 Dev (35B-A3B)Kuaishou · 35B (3.0B active) Q6_K 33.7 GB 3.7 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
88Excellent quality (A+) · 88 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking. Sage-A1.5 35B-A3BInternScience · 36B (3.0B active) Q5_K_M 30.8 GB 4.4 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
87Excellent quality (A+) · 87 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking. Ornith 1.5 35B-A3BDeepReinforce · 36B (3.0B active) Q5_K_M 30.8 GB 4.4 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
95Excellent quality (A+) · 95 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 84/100 (quality, fit, speed and memory use) sets the ranking. Nanbeige4.3 33BBOSS Zhipin (Nanbeige) · 33B Q8_0 43.6 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
86Excellent quality (A+) · 86 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking. Apodex 1.1 Mini (35B-A3B)Apodex · 36B (3.0B active) Q5_K_M 30.8 GB 4.4 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
86Excellent quality (A+) · 86 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking. Ornith 1.0 35B-A3BDeepReinforce · 36B (3.0B active) Q5_K_M 30.8 GB 4.4 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
86Excellent quality (A+) · 86 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking. Agents-A1 35B-A3BInternScience · 35B (3.0B active) Q6_K 34.0 GB 3.7 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
82Great quality (A) · 82 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 83/100 (quality, fit, speed and memory use) sets the ranking. Qwen AgentWorld 35B-A3BAlibaba · 35B (3.0B active) Q4_K_M 28.1 GB 15.2 tok/s Runs well Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Tensor split Distributed Llama Needs the clusterWon’t run on any one of them
90Excellent quality (A+) · 90 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 82/100 (quality, fit, speed and memory use) sets the ranking. Gemma 4 31BGoogle · 31B Q6_K 32.3 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
81Great quality (A) · 81 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 81/100 (quality, fit, speed and memory use) sets the ranking. Laguna XS 2.1 (33B-A3B)Poolside · 33B (3.0B active) Q6_K 33.8 GB 3.7 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
76Great quality (A) · 76 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 80/100 (quality, fit, speed and memory use) sets the ranking. Kolibri 1 (78B-A3.5B)Aleph Alpha · 78B (3.5B active) Q3_K_M 45.2 GB 5.7 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
72Great quality (A) · 72 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Qwen3 Coder Next (80B-A3B)Alibaba · 80B (3.0B active) Q3_K_M 46.0 GB 6.4 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
75Great quality (A) · 75 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. Qwen3-VL 30B-A3B ThinkingAlibaba · 31B (3.3B active) Q6_K 31.3 GB 3.4 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
82Great quality (A) · 82 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 78/100 (quality, fit, speed and memory use) sets the ranking. EXAONE 4.5 33BLG AI Research · 34B Q8_0 43.9 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
72Great quality (A) · 72 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking. Qwen3-Next 80B-A3B ThinkingAlibaba · 81B (3.0B active) Q3_K_M 46.8 GB 6.4 tok/s Slowtight fit Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 77/100 (quality, fit, speed and memory use) sets the ranking. Qwen3-VL 30B-A3B InstructAlibaba · 31B (3.3B active) Q6_K 31.3 GB 3.4 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
70Great quality (A) · 70 of 100: how capable the model is at Q3_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Qwen3 Next 80B-A3B InstructAlibaba · 81B (3.0B active) Q3_K_M 46.8 GB 6.4 tok/s Slowtight fit Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Kimi Linear 48B-A3BMoonshot AI · 49B (3.0B active) Q6_K 47.1 GB 3.7 tok/s Slowtight fit Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
77Great quality (A) · 77 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Seed-OSS 36B InstructByteDance · 36B Q5_K_M 32.9 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
78Great quality (A) · 78 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. QwQ 32BAlibaba · 33B Q8_0 43.6 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
78Great quality (A) · 78 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Qwen3-VL 32B InstructAlibaba · 33B Q8_0 44.3 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
77Great quality (A) · 77 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 76/100 (quality, fit, speed and memory use) sets the ranking. Llama 3.3 Nemotron Super 49B v1.5NVIDIA · 50B Q5_K_M 44.8 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
76Great quality (A) · 76 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Hermes 4.3 36BNous Research · 36B Q5_K_M 32.9 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
63Good quality (B) · 63 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Flint (40B-A2.5B)Recursive · 40B (2.5B active) Q2_K 19.5 GB 10.1 tok/s Runs well Pi / ARM board 1 + Pi / ARM board 2Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
76Great quality (A) · 76 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. OpenReasoning Nemotron 32BNVIDIA · 33B Q8_0 43.6 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
76Great quality (A) · 76 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Skywork OR1 32BSkywork · 33B Q8_0 43.6 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
76Great quality (A) · 76 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Qwen3 32BAlibaba · 33B Q8_0 43.7 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
75Great quality (A) · 75 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Granite 4.2 30BIBM · 29B Q6_K 31.2 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
67Good quality (B) · 67 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Qwen3-Omni 30B-A3B InstructAlibaba · 35B (3.3B active) Q5_K_M 30.9 GB 4.0 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
75Great quality (A) · 75 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 75/100 (quality, fit, speed and memory use) sets the ranking. Olmo 3.1 32B ThinkAi2 · 32B Q6_K 32.2 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
73Great quality (A) · 73 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. InternVL3.5 38BShanghai AI Lab · 38B Q5_K_M 34.5 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
74Great quality (A) · 74 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. EXAONE Deep 32BLG AI Research · 32B Q8_0 42.8 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. GLM-4 32B 0414Zhipu AI · 33B Q6_K 32.3 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
73Great quality (A) · 73 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. GLM-Z1 32B (0414)Zhipu AI · 33B Q6_K 32.3 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
76Great quality (A) · 76 of 100: how capable the model is at Q4_K_M, whatever it runs on. Match score 74/100 (quality, fit, speed and memory use) sets the ranking. Kimi Dev 72BMoonshot AI · 73B Q4_K_M 53.8 GB 0.2 tok/s Very slowtight fit Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
72Great quality (A) · 72 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. DeepSeek R1 Distill Qwen 32BDeepSeek · 33B Q8_0 43.7 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
71Great quality (A) · 71 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. Granite 4.1 30BIBM · 29B Q6_K 30.9 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
71Great quality (A) · 71 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. EXAONE 4.0 32BLG AI Research · 32B Q6_K 32.0 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
71Great quality (A) · 71 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 73/100 (quality, fit, speed and memory use) sets the ranking. Olmo 3.1 32B InstructAi2 · 32B Q6_K 32.2 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. Baichuan M2 32BBaichuan · 33B Q8_0 43.6 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. OpenHands LM 32BAll Hands AI · 33B Q8_0 43.6 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
70Great quality (A) · 70 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. Qwen2.5 32B InstructAlibaba · 33B Q8_0 43.7 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
64Good quality (B) · 64 of 100: how capable the model is at Q2_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. Nemotron 3.5 Bolt 74B-A9BNVIDIA · 75B (9.0B active) Q2_K 35.1 GB 3.0 tok/s Slowtight fit Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
61Good quality (B) · 61 of 100: how capable the model is at Q6_K, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. Axion 1.0 35BAxion · 35B (3.0B active) Q6_K 34.0 GB 3.7 tok/s Slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
67Good quality (B) · 67 of 100: how capable the model is at Q5_K_M, whatever it runs on. Match score 72/100 (quality, fit, speed and memory use) sets the ranking. InternVL3 38BShanghai AI Lab · 38B Q5_K_M 34.5 GB 0.4 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them
68Good quality (B) · 68 of 100: how capable the model is at Q8_0, whatever it runs on. Match score 71/100 (quality, fit, speed and memory use) sets the ranking. Qwen2.5-VL 32B InstructAlibaba · 33B Q8_0 44.4 GB 0.3 tok/s Very slow Pi / ARM board 1 + Pi / ARM board 2 + Pi / ARM board 3 + Pi / ARM board 4Pipeline split llama.cpp RPC Needs the clusterWon’t run on any one of them

Other example clusters

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.

Was this page helpful?