{T}
TokenMath.net
Public Audit HistoryAutomated Pipeline + Manual Verification

AI Pricing Changelog & Provider Audit Trail

Every rate change, prompt caching revision, and context window adjustment on TokenMath is recorded and logged here. Prices are actively verified against first-party provider documentation with per-model verification timestamps.

Total Recorded Audits
6
Continuous daily sync pipeline
Active Provider Feeds
6 Ecosystems
OpenAI, Anthropic, Google, DeepSeek, Meta / Together, Mistral
Verification Policy
First-Party Docs Audit
Cross-verified against official provider rate schedules

Audit History & Rate Modifications

2026-09-12 21:50 UTC
Pipeline Trigger:authoritative first party calibration
✓ Published to Registry3 Rate Adjustments
Calibrated Rates & Specifications:
DeepSeek-V4-FlashDeepSeek
First-Party Documentation Ground Truth
Input: $0.28$0.44/1MOutput: $0.42$1.32/1MCached: $0.005$0.014/1M

Calibrated to active deepseek-v4-flash API specification with tiered peak/off-peak pricing ($0.44/$1.32 peak, $0.22/$0.66 off-peak).

Source: DeepSeek Official API Pricing
GPT-6 AstraOpenAI
First-Party Documentation Ground Truth
Input: $10$10/1MOutput: $50$50/1MCached: $1$1/1M

Updated API ID to canonical gpt-6-astra and promoted lifecycle status to Generally Available (Flagship).

Source: OpenAI Official Documentation
Mistral Small 4Mistral AI
First-Party Documentation Ground Truth
Input: $0.15$0.15/1MOutput: $0.6$0.6/1MCached: $0.015$0.015/1M

Updated canonical API model ID to mistral-small-2603 matching official March 2026 documentation.

Source: Mistral AI Official Documentation
Provider Endpoint Health:
DeepSeekOpenAIMistral AI
2026-09-12 21:40 UTC
Pipeline Trigger:authoritative first party calibration
✓ Published to Registry2 Rate Adjustments
Calibrated Rates & Specifications:
Mistral Large 3Mistral AI
First-Party Documentation Ground Truth
Input: $0.5$0.5/1MOutput: $1.5$1.5/1MCached: $0.1$0.05/1MContext: 128k262k

Calibrated context window to native 256k (262,144 tokens) and prompt cache rate to $0.05/1M (90% discount).

Source: Mistral AI Official Documentation
Mistral Small 4Mistral AI
First-Party Documentation Ground Truth
Input: $0.15$0.15/1MOutput: $0.6$0.6/1MCached: $0.03$0.015/1MContext: 128k262k

Calibrated context window to native 256k (262,144 tokens) and prompt cache rate to $0.015/1M (90% discount).

Source: Mistral AI Official Documentation
Provider Endpoint Health:
Mistral AIGoogle CloudOpenAIAnthropicDeepSeekTogether AI
2026-09-12 21:00 UTC
Pipeline Trigger:first party provider calibration
✓ Published to Registry1 Rate Adjustments
Calibrated Rates & Specifications:
Gemini 3.8 FlashGoogle
First-Party Documentation + Headless Crawler Verification
Input: $0.75$0.75/1MOutput: $3.75$3.75/1MCached: $0.19$0.075/1MContext: 1049k1049k

Context caching rate calibrated to official promotional $0.075/1M tokens through Dec 31, 2026 (subsequently $0.15/1M from Jan 1, 2027) with active promotional cache storage fee of $0.50/1M tokens/hour through Dec 31, 2026 ($1.00/1M/hour standard thereafter).

Source: Google AI Studio & Vertex AI Pricing Documentation
Provider Endpoint Health:
Google CloudOpenAIAnthropicDeepSeekTogether AIMistral AI
2026-09-12 19:48 UTC
Pipeline Trigger:first party provider calibration
✓ Published to Registry7 Rate Adjustments
Calibrated Rates & Specifications:
Claude Opus 5Anthropic
automated + manual cross-verification
Input: $15$5/1MOutput: $75$25/1MCached: $0.5$0.5/1MContext: 200k1000k

Enterprise flagship calibration: $5.00 input / $25.00 output, 1.0M token context window.

Source: Anthropic Official Documentation
Claude Sonnet 5Anthropic
automated + manual cross-verification
Input: $3$2/1MOutput: $15$10/1MCached: $0.2$0.2/1M

Frontier tier price reduction: $2.00 input / $10.00 output.

Source: Anthropic Official Documentation
Claude Fable 5.1Anthropic
manual first-party documentation
Context: 200k1000k

Context window expanded to 1,000,000 tokens for autonomous agentic workflows.

Source: Anthropic Official Documentation
GPT-6 AstraOpenAI
automated + manual cross-verification
Input: $10$10/1MOutput: $30$50/1MCached: $5$1/1MContext: 256k1050k

Frontier reasoning calibration: $10.00 input / $50.00 output, cached input discount reduced to $1.00, context verified at 1.05M tokens.

Source: OpenAI API Official Documentation
GPT-5.6 SolOpenAI
automated + manual cross-verification
Input: $2.5$4/1MOutput: $10$20/1MCached: $2$0.4/1MContext: 200k1050k

Calibrated against current OpenAI pricing: $4.00 input / $20.00 output, $0.40 cached, 1.05M tokens context.

Source: OpenAI API Official Documentation
GPT-5.6 TerraOpenAI
automated + manual cross-verification
Input: $0.15$2/1MOutput: $12$12/1MCached: $1$0.2/1MContext: 128k1050k

Balanced tier rate correction: $2.00 input / $12.00 output, $0.20 cached, 1.05M tokens context.

Source: OpenAI API Official Documentation
GPT-5.6 LunaOpenAI
automated + manual cross-verification
Input: $0.2$0.2/1MOutput: $1.2$1.2/1MCached: $0.1$0.02/1MContext: 128k1050k

Lightweight tier calibration: 90% prompt caching discount ($0.02/M), 1.05M tokens context.

Source: OpenAI API Official Documentation
Provider Endpoint Health:
OpenAIAnthropicGoogle CloudDeepSeekTogether AIMistral AI
2026-09-12 16:36 UTC
Pipeline Trigger:daily heartbeat check
✓ Published to RegistryVerified Clean (No Drift)
All 19 registered frontier models match official provider documentation.
Provider Endpoint Health:
OpenAIAnthropicGoogle CloudDeepSeekTogether AIMistral AI
2026-09-11 20:18 UTC
Pipeline Trigger:automated daily drift sync
✓ Published to Registry8 Rate Adjustments
Calibrated Rates & Specifications:
Claude Opus 5Anthropic
Input: $5$15/1MOutput: $25$75/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

GPT-6 AstraOpenAI
Input: $10$10/1MOutput: $50$30/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

GPT-5.6 SolOpenAI
Input: $4$2.5/1MOutput: $20$10/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

GPT-5.6 TerraOpenAI
Input: $2$0.15/1MOutput: $12$12/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

Gemini 3.8 FlashGoogle
Input: $0.75$0.1/1MOutput: $3.75$0.4/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

DeepSeek-V4.1-FlashDeepSeek
Input: $0.2$0.28/1MOutput: $0.8$0.42/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

Llama 4 Scout (109B)Meta / Open
Input: $0.3$0.1/1MOutput: $0.6$0.32/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

Mistral Small 4Mistral AI
Input: $0.15$0.05/1MOutput: $0.6$0.08/1M

Legacy third-party aggregator feed drift (subsequently overridden by authoritative first-party provider calibration).

Provider Endpoint Health:
OpenAIAnthropicGoogle CloudDeepSeekTogether AIMistral AI

How TokenMath Audits Prices

Our ingestion daemon runs daily cron routines against provider documentation endpoints, OpenRouter models registry, and LiteLLM canonical feeds. Any candidate rate delta exceeding 0.01% triggers automated diff isolation and is cross-referenced with primary provider documentation before entering the canonical registry.