// AI Platform Drift

DeepSeek Flash Latest pricing and output limit changed

OpenRouter model catalogue · published 2026-10-03 21:16 UTC
ACTION NEEDED BUDGET PRICINGLIMITSMODEL
WHAT CHANGEDOn the ~deepseek/deepseek-flash-latest model, prompt pricing increased from $0.0150/M tokens to $0.0200/M tokens (+33.3%), completion pricing decreased from $0.6770/M tokens to $0.4000/M tokens (-40.9%), and max_completion_tokens dropped from 943,718 to 393,216 tokens.
WHY IT MATTERSWorkloads generating more than 393,216 tokens per request will now fail. Cost impact depends on your prompt-to-completion ratio: requests with ten completion tokens for every prompt token see a 40.8% cost reduction, but prompt-heavy workloads see smaller savings or net increases.
WHAT TO DOCheck your logs for any requests to this model exceeding 393,216 completion tokens and either split them or switch models. Reforecast spend based on your actual prompt/completion mix and update any cost alerts.
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: ~deepseek/deepseek-flash-latest
 name: DeepSeek: DeepSeek Flash Latest
-pricing.prompt: 0.000000015
-pricing.completion: 0.000000676999
+pricing.prompt: 0.00000002
+pricing.completion: 0.0000004
 context_length: 1048576
-top_provider.max_completion_tokens: 943718
+top_provider.max_completion_tokens: 393216
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 and token limit changes directly affect engineering budgets and application design decisions.
signalsprompt pricing increased ~33% · completion pricing decreased ~41% · max_completion_tokens reduced by 58% (943718 → 393216)
suppression—
triage modelanthropic/claude-haiku-4.5 ($0.00000)
writer modelanthropic/claude-sonnet-4.5 ($0.00000)