{
 "meta": {
  "generated": "2026-10-16T15:13:49+00:00",
  "fx": {
   "eur_usd": 1.1253,
   "date": "2026-10-16",
   "source": "ECB euro reference rates"
  },
  "metric": {
   "small": "llama.cpp llama-bench, Llama-2 7B Q4_0, tg128 (text generation tok/s); best of FA on/off as posted",
   "large": "gpt-oss-120b MXFP4 (59 GiB weights), text generation tok/s at empty context, single request"
  },
  "license_note": "Benchmark rows cite their sources; llm-speed data CC BY 4.0."
 },
 "items": [
  {
   "id": "mac-studio-m5-max-128",
   "name": "Mac Studio M5 Max 40-core GPU, 128GB / 512GB",
   "family": "Apple",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 5099,
   "price_usd_converted": false,
   "price_eur": 5909,
   "price_eur_converted": false,
   "in_stock": [
    null,
    null
   ],
   "tokens_per_sec": 119.08,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/4167#discussioncomment-17386908",
   "tps_per_1000_usd": 23.35,
   "tps_per_1000_eur": 20.15,
   "large_model_tokens_per_sec": 79.1,
   "large_model_status": null,
   "large_model_source": "https://bestllmfor.com/guides/macbook-pro-m5-max-128gb-ai-benchmark/",
   "large_tps_per_1000_usd": 15.51,
   "large_tps_per_1000_eur": 13.39,
   "kind": "machine",
   "page": "https://localaiprices.com/p/mac-studio-m5-max-128/",
   "history": "https://localaiprices.com/api/v1/history/mac-studio-m5-max-128.json"
  },
  {
   "id": "framework-desktop-128",
   "name": "Framework Desktop 128GB (DIY)",
   "family": "Strix Halo",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 3449,
   "price_usd_converted": false,
   "price_eur": 3889,
   "price_eur_converted": false,
   "in_stock": [
    null,
    null
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 14.77,
   "tps_per_1000_eur": 13.1,
   "large_model_tokens_per_sec": 52.74,
   "large_model_status": null,
   "large_model_source": "https://www.reddit.com/r/LocalLLaMA/comments/1o6u5o4/gptoss20120b_amd_strix_halo_vs_nvidia_dgx_spark/",
   "large_tps_per_1000_usd": 15.29,
   "large_tps_per_1000_eur": 13.56,
   "kind": "machine",
   "page": "https://localaiprices.com/p/framework-desktop-128/",
   "history": "https://localaiprices.com/api/v1/history/framework-desktop-128.json"
  },
  {
   "id": "gmktec-evo-x2-128-1tb",
   "name": "GMKtec EVO-X2 128GB / 1TB",
   "family": "Strix Halo",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 3499.99,
   "price_usd_converted": false,
   "price_eur": 3299.99,
   "price_eur_converted": false,
   "in_stock": [
    true,
    true
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 14.56,
   "tps_per_1000_eur": 15.44,
   "large_model_tokens_per_sec": 52.74,
   "large_model_status": null,
   "large_model_source": "https://www.reddit.com/r/LocalLLaMA/comments/1o6u5o4/gptoss20120b_amd_strix_halo_vs_nvidia_dgx_spark/",
   "large_tps_per_1000_usd": 15.07,
   "large_tps_per_1000_eur": 15.98,
   "kind": "machine",
   "page": "https://localaiprices.com/p/gmktec-evo-x2-128-1tb/",
   "history": "https://localaiprices.com/api/v1/history/gmktec-evo-x2-128-1tb.json"
  },
  {
   "id": "gmktec-evo-x2-128-2tb",
   "name": "GMKtec EVO-X2 128GB / 2TB",
   "family": "Strix Halo",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 3649.99,
   "price_usd_converted": false,
   "price_eur": 3449.99,
   "price_eur_converted": false,
   "in_stock": [
    true,
    true
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 13.96,
   "tps_per_1000_eur": 14.77,
   "large_model_tokens_per_sec": 52.74,
   "large_model_status": null,
   "large_model_source": "https://www.reddit.com/r/LocalLLaMA/comments/1o6u5o4/gptoss20120b_amd_strix_halo_vs_nvidia_dgx_spark/",
   "large_tps_per_1000_usd": 14.45,
   "large_tps_per_1000_eur": 15.29,
   "kind": "machine",
   "page": "https://localaiprices.com/p/gmktec-evo-x2-128-2tb/",
   "history": "https://localaiprices.com/api/v1/history/gmktec-evo-x2-128-2tb.json"
  },
  {
   "id": "gmktec-evo-x3-128-2tb",
   "name": "GMKtec EVO-X3 128GB / 2TB",
   "family": "Strix Halo",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 3799.99,
   "price_usd_converted": false,
   "price_eur": 3549.99,
   "price_eur_converted": false,
   "in_stock": [
    true,
    true
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 13.41,
   "tps_per_1000_eur": 14.35,
   "large_model_tokens_per_sec": 52.74,
   "large_model_status": null,
   "large_model_source": "https://www.reddit.com/r/LocalLLaMA/comments/1o6u5o4/gptoss20120b_amd_strix_halo_vs_nvidia_dgx_spark/",
   "large_tps_per_1000_usd": 13.88,
   "large_tps_per_1000_eur": 14.86,
   "kind": "machine",
   "page": "https://localaiprices.com/p/gmktec-evo-x3-128-2tb/",
   "history": "https://localaiprices.com/api/v1/history/gmktec-evo-x3-128-2tb.json"
  },
  {
   "id": "minisforum-ms-s1-max-128",
   "name": "Minisforum MS-S1 MAX 128GB / 2TB",
   "family": "Strix Halo",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 3879.0,
   "price_usd_converted": false,
   "price_eur": 3447.08,
   "price_eur_converted": true,
   "in_stock": [
    true
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 13.13,
   "tps_per_1000_eur": 14.78,
   "large_model_tokens_per_sec": 52.74,
   "large_model_status": null,
   "large_model_source": "https://www.reddit.com/r/LocalLLaMA/comments/1o6u5o4/gptoss20120b_amd_strix_halo_vs_nvidia_dgx_spark/",
   "large_tps_per_1000_usd": 13.6,
   "large_tps_per_1000_eur": 15.3,
   "kind": "machine",
   "page": "https://localaiprices.com/p/minisforum-ms-s1-max-128/",
   "history": "https://localaiprices.com/api/v1/history/minisforum-ms-s1-max-128.json"
  },
  {
   "id": "nvidia-dgx-spark",
   "name": "NVIDIA DGX Spark (Founders Edition)",
   "family": "DGX Spark / GB10",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 4699,
   "price_usd_converted": false,
   "price_eur": 4175.78,
   "price_eur_converted": true,
   "in_stock": [
    null
   ],
   "tokens_per_sec": 57.21,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-14687095",
   "tps_per_1000_usd": 12.17,
   "tps_per_1000_eur": 13.7,
   "large_model_tokens_per_sec": 58.72,
   "large_model_status": null,
   "large_model_source": "https://github.com/ggml-org/llama.cpp/blob/master/benches/dgx-spark/dgx-spark.md",
   "large_tps_per_1000_usd": 12.5,
   "large_tps_per_1000_eur": 14.06,
   "kind": "machine",
   "page": "https://localaiprices.com/p/nvidia-dgx-spark/",
   "history": "https://localaiprices.com/api/v1/history/nvidia-dgx-spark.json"
  },
  {
   "id": "beelink-gtr9-pro-128",
   "name": "Beelink GTR9 Pro 128GB / 2TB",
   "family": "Strix Halo",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 4349.0,
   "price_usd_converted": false,
   "price_eur": 3864.75,
   "price_eur_converted": true,
   "in_stock": [
    true,
    true
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 11.72,
   "tps_per_1000_eur": 13.18,
   "large_model_tokens_per_sec": 52.74,
   "large_model_status": null,
   "large_model_source": "https://www.reddit.com/r/LocalLLaMA/comments/1o6u5o4/gptoss20120b_amd_strix_halo_vs_nvidia_dgx_spark/",
   "large_tps_per_1000_usd": 12.13,
   "large_tps_per_1000_eur": 13.65,
   "kind": "machine",
   "page": "https://localaiprices.com/p/beelink-gtr9-pro-128/",
   "history": "https://localaiprices.com/api/v1/history/beelink-gtr9-pro-128.json"
  },
  {
   "id": "mac-studio-m5-ultra-80-256",
   "name": "Mac Studio M5 Ultra 80-core GPU, 256GB / 1TB",
   "family": "Apple",
   "memory_gb": 256,
   "condition": "new",
   "price_usd": 10799,
   "price_usd_converted": false,
   "price_eur": 12529,
   "price_eur_converted": false,
   "in_stock": [
    null,
    null
   ],
   "tokens_per_sec": 179.1,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/4167#discussioncomment-18683605",
   "tps_per_1000_usd": 16.58,
   "tps_per_1000_eur": 14.29,
   "large_model_tokens_per_sec": 130,
   "large_model_status": null,
   "large_model_source": "https://www.youtube.com/watch?v=c_58D7ixOQI",
   "large_tps_per_1000_usd": 12.04,
   "large_tps_per_1000_eur": 10.38,
   "kind": "machine",
   "page": "https://localaiprices.com/p/mac-studio-m5-ultra-80-256/",
   "history": "https://localaiprices.com/api/v1/history/mac-studio-m5-ultra-80-256.json"
  },
  {
   "id": "macbook-pro-16-m5-max-128",
   "name": "MacBook Pro 16\" M5 Max 40-core GPU, 128GB / 2TB",
   "family": "Apple",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 6999,
   "price_usd_converted": false,
   "price_eur": 7999,
   "price_eur_converted": false,
   "in_stock": [
    null,
    null
   ],
   "tokens_per_sec": 119.08,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/4167#discussioncomment-17386908",
   "tps_per_1000_usd": 17.01,
   "tps_per_1000_eur": 14.89,
   "large_model_tokens_per_sec": 79.1,
   "large_model_status": null,
   "large_model_source": "https://bestllmfor.com/guides/macbook-pro-m5-max-128gb-ai-benchmark/",
   "large_tps_per_1000_usd": 11.3,
   "large_tps_per_1000_eur": 9.89,
   "kind": "machine",
   "page": "https://localaiprices.com/p/macbook-pro-16-m5-max-128/",
   "history": "https://localaiprices.com/api/v1/history/macbook-pro-16-m5-max-128.json"
  },
  {
   "id": "dell-pro-max-gb10",
   "name": "Dell Pro Max / Pro Precision with GB10, 128GB / 2TB",
   "family": "DGX Spark / GB10",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 5903,
   "price_usd_converted": false,
   "price_eur": 5245.71,
   "price_eur_converted": true,
   "in_stock": [
    null
   ],
   "tokens_per_sec": 57.21,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-14687095",
   "tps_per_1000_usd": 9.69,
   "tps_per_1000_eur": 10.91,
   "large_model_tokens_per_sec": 58.72,
   "large_model_status": null,
   "large_model_source": "https://github.com/ggml-org/llama.cpp/blob/master/benches/dgx-spark/dgx-spark.md",
   "large_tps_per_1000_usd": 9.95,
   "large_tps_per_1000_eur": 11.19,
   "kind": "machine",
   "page": "https://localaiprices.com/p/dell-pro-max-gb10/",
   "history": "https://localaiprices.com/api/v1/history/dell-pro-max-gb10.json"
  },
  {
   "id": "asus-ascent-gx10",
   "name": "ASUS Ascent GX10 (GB10)",
   "family": "DGX Spark / GB10",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": 5999,
   "price_usd_converted": false,
   "price_eur": 5331.02,
   "price_eur_converted": true,
   "in_stock": [
    null
   ],
   "tokens_per_sec": 57.21,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-14687095",
   "tps_per_1000_usd": 9.54,
   "tps_per_1000_eur": 10.73,
   "large_model_tokens_per_sec": 58.72,
   "large_model_status": null,
   "large_model_source": "https://github.com/ggml-org/llama.cpp/blob/master/benches/dgx-spark/dgx-spark.md",
   "large_tps_per_1000_usd": 9.79,
   "large_tps_per_1000_eur": 11.01,
   "kind": "machine",
   "page": "https://localaiprices.com/p/asus-ascent-gx10/",
   "history": "https://localaiprices.com/api/v1/history/asus-ascent-gx10.json"
  },
  {
   "id": "gmktec-evo-x2-64-1tb",
   "name": "GMKtec EVO-X2 64GB / 1TB",
   "family": "Strix Halo",
   "memory_gb": 64,
   "condition": "new",
   "price_usd": 2199.99,
   "price_usd_converted": false,
   "price_eur": 1999.99,
   "price_eur_converted": false,
   "in_stock": [
    true,
    true
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 23.16,
   "tps_per_1000_eur": 25.48,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/gmktec-evo-x2-64-1tb/",
   "history": "https://localaiprices.com/api/v1/history/gmktec-evo-x2-64-1tb.json"
  },
  {
   "id": "minisforum-ms-s1-max-64",
   "name": "Minisforum MS-S1 MAX 64GB / 2TB",
   "family": "Strix Halo",
   "memory_gb": 64,
   "condition": "new",
   "price_usd": 2599.0,
   "price_usd_converted": false,
   "price_eur": 2309.61,
   "price_eur_converted": true,
   "in_stock": [
    true
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 19.6,
   "tps_per_1000_eur": 22.06,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/minisforum-ms-s1-max-64/",
   "history": "https://localaiprices.com/api/v1/history/minisforum-ms-s1-max-64.json"
  },
  {
   "id": "framework-desktop-64",
   "name": "Framework Desktop 64GB (DIY)",
   "family": "Strix Halo",
   "memory_gb": 64,
   "condition": "new",
   "price_usd": 2485.79,
   "price_usd_converted": true,
   "price_eur": 2209,
   "price_eur_converted": false,
   "in_stock": [
    null
   ],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": 20.5,
   "tps_per_1000_eur": 23.06,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/framework-desktop-64/",
   "history": "https://localaiprices.com/api/v1/history/framework-desktop-64.json"
  },
  {
   "id": "hp-z2-mini-g1a",
   "name": "HP Z2 Mini G1a (Ryzen AI Max+ PRO 395)",
   "family": "Strix Halo",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": null,
   "price_usd_converted": null,
   "price_eur": null,
   "price_eur_converted": null,
   "in_stock": [],
   "tokens_per_sec": 50.95,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15671492",
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": 52.74,
   "large_model_status": null,
   "large_model_source": "https://www.reddit.com/r/LocalLLaMA/comments/1o6u5o4/gptoss20120b_amd_strix_halo_vs_nvidia_dgx_spark/",
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/hp-z2-mini-g1a/",
   "history": "https://localaiprices.com/api/v1/history/hp-z2-mini-g1a.json"
  },
  {
   "id": "msi-edgexpert",
   "name": "MSI EdgeXpert MS-C931 (GB10)",
   "family": "DGX Spark / GB10",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": null,
   "price_usd_converted": null,
   "price_eur": null,
   "price_eur_converted": null,
   "in_stock": [],
   "tokens_per_sec": 57.21,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-14687095",
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": 58.72,
   "large_model_status": null,
   "large_model_source": "https://github.com/ggml-org/llama.cpp/blob/master/benches/dgx-spark/dgx-spark.md",
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/msi-edgexpert/",
   "history": "https://localaiprices.com/api/v1/history/msi-edgexpert.json"
  },
  {
   "id": "lenovo-thinkstation-pgx",
   "name": "Lenovo ThinkStation PGX (GB10)",
   "family": "DGX Spark / GB10",
   "memory_gb": 128,
   "condition": "new",
   "price_usd": null,
   "price_usd_converted": null,
   "price_eur": null,
   "price_eur_converted": null,
   "in_stock": [],
   "tokens_per_sec": 57.21,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-14687095",
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": 58.72,
   "large_model_status": null,
   "large_model_source": "https://github.com/ggml-org/llama.cpp/blob/master/benches/dgx-spark/dgx-spark.md",
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/lenovo-thinkstation-pgx/",
   "history": "https://localaiprices.com/api/v1/history/lenovo-thinkstation-pgx.json"
  },
  {
   "id": "mac-studio-m5-ultra-96",
   "name": "Mac Studio M5 Ultra 64-core GPU, 96GB / 1TB",
   "family": "Apple",
   "memory_gb": 96,
   "condition": "new",
   "price_usd": 5499,
   "price_usd_converted": false,
   "price_eur": 6699,
   "price_eur_converted": false,
   "in_stock": [
    null,
    null
   ],
   "tokens_per_sec": null,
   "benchmark_source": null,
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": null,
   "large_model_status": "no public result yet",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/mac-studio-m5-ultra-96/",
   "history": "https://localaiprices.com/api/v1/history/mac-studio-m5-ultra-96.json"
  },
  {
   "id": "mac-mini-m5-pro-16",
   "name": "Mac mini M5 Pro 16-core GPU, 64GB / 512GB",
   "family": "Apple",
   "memory_gb": 64,
   "condition": "new",
   "price_usd": 2699,
   "price_usd_converted": false,
   "price_eur": 3129,
   "price_eur_converted": false,
   "in_stock": [
    null,
    null
   ],
   "tokens_per_sec": null,
   "benchmark_source": null,
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/mac-mini-m5-pro-16/",
   "history": "https://localaiprices.com/api/v1/history/mac-mini-m5-pro-16.json"
  },
  {
   "id": "mac-mini-m5-pro-20",
   "name": "Mac mini M5 Pro 20-core GPU, 64GB / 512GB",
   "family": "Apple",
   "memory_gb": 64,
   "condition": "new",
   "price_usd": 2899,
   "price_usd_converted": false,
   "price_eur": 3349,
   "price_eur_converted": false,
   "in_stock": [
    null,
    null
   ],
   "tokens_per_sec": null,
   "benchmark_source": null,
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "machine",
   "page": "https://localaiprices.com/p/mac-mini-m5-pro-20/",
   "history": "https://localaiprices.com/api/v1/history/mac-mini-m5-pro-20.json"
  },
  {
   "id": "rtx-5090",
   "name": "NVIDIA RTX 5090 32GB (new)",
   "family": "GPU",
   "memory_gb": 32,
   "condition": "new",
   "price_usd": null,
   "price_usd_converted": null,
   "price_eur": null,
   "price_eur_converted": null,
   "in_stock": [],
   "tokens_per_sec": 314.0,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-15160999",
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "gpu-card",
   "page": "https://localaiprices.com/p/rtx-5090/",
   "history": "https://localaiprices.com/api/v1/history/rtx-5090.json"
  },
  {
   "id": "rtx-4090-used",
   "name": "NVIDIA RTX 4090 24GB (used)",
   "family": "GPU",
   "memory_gb": 24,
   "condition": "used",
   "price_usd": null,
   "price_usd_converted": null,
   "price_eur": null,
   "price_eur_converted": null,
   "in_stock": [],
   "tokens_per_sec": 191.57,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-14200008",
   "tps_per_1000_usd": null,
   "tps_per_1000_eur": null,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "gpu-card",
   "page": "https://localaiprices.com/p/rtx-4090-used/",
   "history": "https://localaiprices.com/api/v1/history/rtx-4090-used.json"
  },
  {
   "id": "rtx-3090-used",
   "name": "NVIDIA RTX 3090 24GB (used)",
   "family": "GPU",
   "memory_gb": 24,
   "condition": "used",
   "price_usd": 700,
   "price_usd_converted": false,
   "price_eur": 622.06,
   "price_eur_converted": true,
   "in_stock": [
    null
   ],
   "tokens_per_sec": 172.26,
   "benchmark_source": "https://github.com/ggml-org/llama.cpp/discussions/15013#discussioncomment-13963584",
   "tps_per_1000_usd": 246.09,
   "tps_per_1000_eur": 276.92,
   "large_model_tokens_per_sec": null,
   "large_model_status": "doesn't fit in memory",
   "large_model_source": null,
   "large_tps_per_1000_usd": null,
   "large_tps_per_1000_eur": null,
   "kind": "gpu-card",
   "page": "https://localaiprices.com/p/rtx-3090-used/",
   "history": "https://localaiprices.com/api/v1/history/rtx-3090-used.json"
  }
 ]
}