OpenRouter provider table for meta-llama/llama-3.3-70b-instruct, re-read 2026-10-17 (per-provider price, P50 latency, throughput, uptime)

Fetched 2026-10-17, same URL as svc-src-openrouter-provider-table-2026-10-16. Per-provider rows for Llama 3.3 70B Instruct, P50 over OpenRouter's own routed traffic. DeepInfra (Turbo) $0.100/$0.320, 0.62s latency, 19 tps, 99.94% uptime (was 0.48s / 16 tps / 98.80% on 2026-10-16). NovitaAI $0.135/$0.400, 1.17s, 45 tps, 99.15% (was 1.05s / 44 tps / 98.40%). AkashML $0.200/$0.520 (cache read $0.100), 0.97s, 37 tps, 99.83%. Parasail $0.220/$0.500 (cache read $0.110), 0.60s, 55 tps, 100.00%. Cloudflare $0.293/$2.253, 0.63s, 45 tps, 99.34%. SambaNova Turbo list $0.60/$1.20 struck through next to a 25% off badge rendering $0.450/$0.900, 1.01s, 93 tps, 99.90% (was 0.81s / 104 tps / 99.83%). Groq $0.590/$0.790 (cache read $0.295), 0.50s, 198 tps, 100.00% (was 0.52s / 193 tps). CoreWeave $0.710/$0.710 (cache read $0.710), 0.13s, 86 tps, 100.00% (was 0.12s / 85 tps). Google Vertex (US flag) $0.720/$0.720, 1.19s, 92 tps; a second Vertex row at the same price has no latency cell. Together $1.040/$1.040, 0.94s, 52 tps, 95.75% — the 2026-10-16 read had no latency cell for Together. Page headline price is still the floor route $0.10/$0.32. Best-across-providers throughput widget reads 198 tok/s (Groq); best latency widget reads 0.13s (CoreWeave). List prices on every row are unchanged from 2026-10-16 except the SambaNova discount now renders as a struck-through list price rather than a badge next to the full rate. These are OpenRouter-measured P50 numbers over its own traffic, not vendor claims and not the Artificial Analysis medians stored on the individual provider records.