{
  "model": "meta-llama/llama-3.3-70b-instruct",
  "date": "2026-10-07",
  "suite": "gsm8k-200",
  "items": 200,
  "seeds": 1,
  "max_tokens": 384,
  "latency_requests": 20,
  "spent_usd": 1.13,
  "providers": [
    {
      "provider": "DeepInfra",
      "tag": "deepinfra/turbo",
      "quantization": "fp8",
      "context_length": 131072,
      "price_in_per_m": 0.09999999999999999,
      "price_out_per_m": 0.32,
      "served_by": "DeepInfra",
      "accuracy_pct": 98.0,
      "accuracy_per_seed": [
        98.0
      ],
      "accuracy_errors": 0,
      "lat_requests": 20,
      "lat_ok": 20,
      "lat_errors": [],
      "lat_ttft_p50_ms": 513.7575617991388,
      "lat_ttft_p90_ms": 1298.0947456788272,
      "lat_tpot_p50_ms": 51.935950544797024,
      "lat_tpot_p90_ms": 77.05908308951801,
      "lat_decode_tok_s_p50": 19.293897992443256,
      "rerun": false
    },
    {
      "provider": "Novita",
      "tag": "novita/bf16",
      "quantization": "bf16",
      "context_length": 12288,
      "price_in_per_m": 0.135,
      "price_out_per_m": 0.39999999999999997,
      "served_by": "Novita",
      "accuracy_pct": 94.5,
      "accuracy_per_seed": [
        94.5
      ],
      "accuracy_errors": 1,
      "lat_requests": 20,
      "lat_ok": 18,
      "lat_errors": [
        "HTTP 429: {\"error\":{\"message\":\"Provider returned error\",\"code\":429,\"metadata\":{\"raw\":\"meta-llama/llama-3.3-70b-instruct is temporarily rate-limited upstream. Please retry shortly, or add your own key to accumul"
      ],
      "lat_ttft_p50_ms": 765.9394375514239,
      "lat_ttft_p90_ms": 1145.6701452843845,
      "lat_tpot_p50_ms": 19.174966541561105,
      "lat_tpot_p90_ms": 23.795805178610063,
      "lat_decode_tok_s_p50": 52.15133018326757,
      "rerun": true
    },
    {
      "provider": "AkashML",
      "tag": "akashml/fp8",
      "quantization": "fp8",
      "context_length": 131072,
      "price_in_per_m": 0.19999999999999998,
      "price_out_per_m": 0.52,
      "served_by": "AkashML",
      "accuracy_pct": 96.5,
      "accuracy_per_seed": [
        96.5
      ],
      "accuracy_errors": 0,
      "lat_requests": 20,
      "lat_ok": 13,
      "lat_errors": [
        "HTTP 429: {\"error\":{\"message\":\"Provider returned error\",\"code\":429,\"metadata\":{\"raw\":\"meta-llama/llama-3.3-70b-instruct is temporarily rate-limited upstream. Please retry shortly, or add your own key to accumul"
      ],
      "lat_ttft_p50_ms": 732.1795825846493,
      "lat_ttft_p90_ms": 1682.5929751619697,
      "lat_tpot_p50_ms": 38.00621351939604,
      "lat_tpot_p90_ms": 63.15194608095813,
      "lat_decode_tok_s_p50": 26.31148718589557,
      "rerun": true
    },
    {
      "provider": "Parasail",
      "tag": "parasail/fp8",
      "quantization": "fp8",
      "context_length": 131072,
      "price_in_per_m": 0.22,
      "price_out_per_m": 0.5,
      "served_by": "Parasail",
      "accuracy_pct": 95.5,
      "accuracy_per_seed": [
        95.5
      ],
      "accuracy_errors": 0,
      "lat_requests": 20,
      "lat_ok": 20,
      "lat_errors": [],
      "lat_ttft_p50_ms": 624.3706876412034,
      "lat_ttft_p90_ms": 1077.2220672108233,
      "lat_tpot_p50_ms": 15.027983964658624,
      "lat_tpot_p90_ms": 29.527066628949548,
      "lat_decode_tok_s_p50": 66.54714090446463,
      "rerun": true
    },
    {
      "provider": "Cloudflare",
      "tag": "cloudflare/fp8",
      "quantization": "fp8",
      "context_length": 24000,
      "price_in_per_m": 0.293,
      "price_out_per_m": 2.2529999999999997,
      "served_by": "Cloudflare",
      "accuracy_pct": 94.0,
      "accuracy_per_seed": [
        94.0
      ],
      "accuracy_errors": 3,
      "lat_requests": 20,
      "lat_ok": 20,
      "lat_errors": [],
      "lat_ttft_p50_ms": 742.6155626308173,
      "lat_ttft_p90_ms": 1452.915521292016,
      "lat_tpot_p50_ms": 15.76800800673892,
      "lat_tpot_p90_ms": 31.469115631554423,
      "lat_decode_tok_s_p50": 63.54094868149187,
      "rerun": true
    },
    {
      "provider": "SambaNova",
      "tag": "sambanova-turbo",
      "quantization": "unknown",
      "context_length": 131072,
      "price_in_per_m": 0.44999999999999996,
      "price_out_per_m": 0.8999999999999999,
      "served_by": "SambaNova",
      "accuracy_pct": 96.5,
      "accuracy_per_seed": [
        96.5
      ],
      "accuracy_errors": 0,
      "lat_requests": 20,
      "lat_ok": 20,
      "lat_errors": [],
      "lat_ttft_p50_ms": 428.4979784861207,
      "lat_ttft_p90_ms": 960.0732420571148,
      "lat_tpot_p50_ms": 3.4792037589438993,
      "lat_tpot_p90_ms": 4.050756030551651,
      "lat_decode_tok_s_p50": 287.43280738438966,
      "rerun": false
    },
    {
      "provider": "Groq",
      "tag": "groq",
      "quantization": "unknown",
      "context_length": 131072,
      "price_in_per_m": 0.59,
      "price_out_per_m": 0.7899999999999999,
      "served_by": "Groq",
      "accuracy_pct": 95.5,
      "accuracy_per_seed": [
        95.5
      ],
      "accuracy_errors": 0,
      "lat_requests": 20,
      "lat_ok": 20,
      "lat_errors": [],
      "lat_ttft_p50_ms": 330.0232086330652,
      "lat_ttft_p90_ms": 468.6883466783911,
      "lat_tpot_p50_ms": 2.878454278426667,
      "lat_tpot_p90_ms": 3.0569419106133062,
      "lat_decode_tok_s_p50": 347.4087839895391,
      "rerun": true
    },
    {
      "provider": "CoreWeave",
      "tag": "coreweave/fp16",
      "quantization": "fp16",
      "context_length": 128000,
      "price_in_per_m": 0.71,
      "price_out_per_m": 0.71,
      "served_by": "CoreWeave",
      "accuracy_pct": 95.0,
      "accuracy_per_seed": [
        95.0
      ],
      "accuracy_errors": 0,
      "lat_requests": 20,
      "lat_ok": 20,
      "lat_errors": [],
      "lat_ttft_p50_ms": 382.16829183511436,
      "lat_ttft_p90_ms": 830.9694997500628,
      "lat_tpot_p50_ms": 12.15985088950455,
      "lat_tpot_p90_ms": 12.605411553499744,
      "lat_decode_tok_s_p50": 82.23999361577701,
      "rerun": false
    },
    {
      "provider": "Google",
      "tag": "google-vertex/us-central1",
      "quantization": "unknown",
      "context_length": 128000,
      "price_in_per_m": 0.72,
      "price_out_per_m": 0.72,
      "error": "RuntimeError: HTTP 404: {\"message\": \"No endpoints found for meta-llama/llama-3.3-70b-instruct. Every candidate endpoint was removed during routing: Filter by Allowed Providers removed deepinfra/turbo, novita/bf16, akashml/fp",
      "rerun": false
    },
    {
      "provider": "Google",
      "tag": "google-vertex",
      "quantization": "unknown",
      "context_length": 128000,
      "price_in_per_m": 0.72,
      "price_out_per_m": 0.72,
      "error": "RuntimeError: HTTP 404: {\"message\": \"No endpoints found for meta-llama/llama-3.3-70b-instruct. Every candidate endpoint was removed during routing: Filter by Allowed Providers removed deepinfra/turbo, novita/bf16, akashml/fp",
      "rerun": false
    },
    {
      "provider": "Together",
      "tag": "together",
      "quantization": "unknown",
      "context_length": 131072,
      "price_in_per_m": 1.04,
      "price_out_per_m": 1.04,
      "served_by": "Together",
      "accuracy_pct": 95.5,
      "accuracy_per_seed": [
        95.5
      ],
      "accuracy_errors": 0,
      "lat_requests": 20,
      "lat_ok": 20,
      "lat_errors": [],
      "lat_ttft_p50_ms": 642.7210834808648,
      "lat_ttft_p90_ms": 1993.7644044402987,
      "lat_tpot_p50_ms": 14.617510557827922,
      "lat_tpot_p90_ms": 21.533369503604867,
      "lat_decode_tok_s_p50": 68.42930606130838,
      "rerun": false
    }
  ],
  "notes": "Providers first run with request failures (Cloudflare rejects `seed`; rate limits elsewhere) were re-run with retries and a seed fallback; failures are reported separately."
}