// AI Platform Drift

Llama 3.3 70B prompt price up 168%, completion up 58%

OpenRouter model catalogue · published 2026-10-13 00:42 UTC
ACTION NEEDED BUDGET PRICINGMODELLIMITS
WHAT CHANGEDOpenRouter raised pricing for meta-llama/llama-3.3-70b-instruct effective 13 October 2026. Captured completion base rate changed from $0.50 to $0.79 per 1M tokens. Captured prompt base rate changed from $0.22 to $0.59 per 1M tokens. Maximum completion tokens doubled from 16,384 to 32,768.
WHY IT MATTERSUsing the captured base rates for meta-llama/llama-3.3-70b-instruct, 100M completion tokens per day cost $50.00 before and $79.00 after, excluding all other charges. Using the captured base rates for meta-llama/llama-3.3-70b-instruct, 100M prompt tokens per day cost $22.00 before and $59.00 after, excluding all other charges. Workloads with equal prompt and completion volume see a 91.7% total cost increase; workloads generating ten completion tokens for every prompt token see a 62.6% increase.
WHAT TO DOReforecast your OpenRouter spend for this model using the new rates and notify whoever owns the budget. To estimate the added completion cost for meta-llama/llama-3.3-70b-instruct in dollars, multiply completion token volume by 0.29 and divide by 1,000,000, using the captured base-rate difference in dollars per 1M tokens. To estimate the added prompt cost for meta-llama/llama-3.3-70b-instruct in dollars, multiply prompt token volume by 0.37 and divide by 1,000,000, using the captured base-rate difference in dollars per 1M tokens.
Model-reported confidence 95% (not measured accuracy) Primary source ↗ written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
 id: meta-llama/llama-3.3-70b-instruct
 name: Meta: Llama 3.3 70B Instruct
-pricing.prompt: 0.00000022
-pricing.completion: 0.0000005
+pricing.prompt: 0.00000059
+pricing.completion: 0.00000079
 context_length: 131072
-top_provider.max_completion_tokens: 16384
+top_provider.max_completion_tokens: 32768
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
triage verdictMATERIAL at 0.98, needed 0.5
triage reasoningPricing increased significantly on both prompt and completion tokens; max output tokens doubled.
signalspricing.prompt increased 2.68x (0.00000022 → 0.00000059 USD/token) · pricing.completion increased 1.58x (0.0000005 → 0.00000079 USD/token) · max_completion_tokens doubled (16384 → 32768)
suppression—
triage modelanthropic/claude-haiku-4.5 ($0.00000)
writer modelanthropic/claude-sonnet-4.5 ($0.00000)