Skip to content
llm-spend
GitHub
Provenance

Freshness and provenance command center

How current the last full catalog audit is, how much of it is officially published versus derived or estimated, what changes on a schedule and when, and where every rate traces back to.

Trust pulse

Catalog freshness at a glance

68 tracked lanes, audited together as of 30 Aug UTC. Every figure below is read directly from catalog metadata — freshness never changes a lane’s published confidence.

3d ago
Last full catalog audit
current
Freshness status · 30 Aug UTC
3
Lanes with a scheduled rate change ahead
4 / 68
Lanes with no cache meter
Input / output confidence
official67derived0estimate1
68 lanes counted
Cached-input confidence
official57derived6estimate1
64 lanes counted · 4 lanes excluded — no cache meter
Fixed thresholds

Current: audited 0–7 days ago. Review due: 830 days ago. Stale: more than 30 days ago. These bands apply to the whole catalog at once (one audit sweep covers every lane) and never change a lane’s published confidence — a stale audit means the evidence needs a closer look, not that the model or its price is worse.

Filters

Focus the trust audit

68 of 68 lanesaudited 30 Aug UTC
Lane records

Every tracked lane, provenance first

68 lanes total. Confidence and cache-meter columns are read straight from the catalog; the “Upcoming” column names lanes with at least one future rate boundary — see the timeline below for the details.

ProviderModelDeploymentInput confidenceCached confidenceEffectiveUpcoming
KimiKimi K3
Kimi direct API
Direct APIofficialofficial2026-07-17
KimiKimi K2.6Foundry ·Globalofficialofficial2026-07-23
KimiKimi K2.7 CodeFoundry ·Globalofficialofficial2026-07-11
KimiKimi K2.5 ThinkingFoundry ·Globalofficialofficial2026-07-23
KimiKimi K3
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-08-21
KimiKimi K2.6
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-07-11
DeepSeekDeepSeek-V4 ProFoundry ·Globalofficialofficial2026-07-11
DeepSeekDeepSeek-V4 FlashFoundry ·Globalofficialofficial2026-07-11
DeepSeekDeepSeek-V4 Pro
DeepSeek direct API
Direct APIofficialofficial2026-07-11
DeepSeekDeepSeek-V4 Flash
DeepSeek direct API
Direct APIofficialofficial2026-07-11
DeepSeekDeepSeek-V4 Pro
Fireworks direct API
Direct APIofficialno cache meter2026-07-11
DeepSeekDeepSeek-V4 ProFoundry ·Data Zoneofficialofficial2026-07-27
DeepSeekDeepSeek-V4 Pro
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-07-11
DeepSeekDeepSeek-V4 FlashFoundry ·Data Zoneofficialofficial2026-08-22
DeepSeekDeepSeek-V4 Flash
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-08-22
GLMGLM 5
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-06-01
GLMGLM 5.1
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-07-11
GLMGLM 5.2
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-07-01
GLMGLM-5
Z.ai direct API
Direct APIofficialofficial2026-07-11
GLMGLM-5-Turbo
Z.ai direct API
Direct APIofficialofficial2026-07-29
GLMGLM-5.1
Z.ai direct API
Direct APIofficialofficial2026-07-26
GLMGLM-5.2
Z.ai direct API
Direct APIofficialofficial2026-07-26
GLMGLM-5.3
Z.ai direct API
Direct APIofficialofficial2026-08-19
GLMGLM-5.3-Flash
Z.ai direct API
Direct APIofficialofficial2026-08-28scheduled ↓
OpenAI / Azure OpenAIGPT-5.6 SolFoundry ·Globalofficialofficial2026-07-21
OpenAI / Azure OpenAIGPT-5.6 SolFoundry ·Data Zoneofficialofficial2026-07-21
OpenAI / Azure OpenAIGPT-5.6 Sol Long ContextFoundry ·Globalofficialofficial2026-07-21
OpenAI / Azure OpenAIGPT-5.6 TerraFoundry ·Globalofficialofficial2026-08-20
OpenAI / Azure OpenAIGPT-5.6 Terra Long ContextFoundry ·Globalofficialofficial2026-08-20
OpenAI / Azure OpenAIGPT-5.6 LunaFoundry ·Globalofficialofficial2026-08-20
OpenAI / Azure OpenAIGPT-5.6 Luna Long ContextFoundry ·Globalofficialofficial2026-08-20
OpenAI / Azure OpenAIGPT-5.5Foundry ·Globalofficialofficial2026-07-11
OpenAI / Azure OpenAIGPT-5.5Foundry ·Data Zoneofficialofficial2026-07-11
OpenAI / Azure OpenAIGPT-5.5 Long ContextFoundry ·Globalofficialofficial2026-07-11
OpenAI / Azure OpenAIGPT-5.3 Codex / ChatFoundry ·Globalofficialofficial2026-07-11
OpenAI / Azure OpenAIGPT-5.2 / CodexFoundry ·Globalofficialofficial2026-07-11
OpenAI / Azure OpenAIGPT-5.2 / CodexFoundry ·Data Zoneofficialofficial2026-07-11
ClaudeClaude Opus 5Direct APIofficialofficial2026-07-25
ClaudeClaude Opus 4.8 (Standard)Direct APIofficialofficial2026-07-11
ClaudeClaude Sonnet 5Direct APIofficialofficial2026-07-11
ClaudeClaude Sonnet 5Foundry ·Globalestimateestimate2026-07-11
ClaudeClaude Fable 5.1Direct APIofficialofficial2026-09-01
ClaudeClaude Mythos 5.1Direct APIofficialofficial2026-09-01
ClaudeClaude Fable 5Direct APIofficialofficial2026-07-11
ClaudeClaude Mythos 5Direct APIofficialofficial2026-07-11
GeminiGemini 3.7 FlashFoundry ·Globalofficialofficial2026-08-14scheduled ↓
GeminiGemini 3.6 FlashFoundry ·Globalofficialofficial2026-08-14scheduled ↓
GeminiGemini 3.5 FlashFoundry ·Globalofficialofficial2026-07-21
GrokGrok 4.6
xAI direct API
Direct APIofficialofficial2026-08-12
GrokGrok 4.5
xAI direct API
Direct APIofficialofficial2026-07-21
GrokGrok-4.3Foundry ·Globalofficialofficial2026-07-23
GrokGrok-4.3Foundry ·Data Zoneofficialofficial2026-07-26
GrokGrok 4.1 FastFoundry ·Globalofficialno cache meter2026-07-19
QwenQwen3.8 Max
Model Studio (Intl)
Direct APIofficialofficial2026-08-04
QwenQwen3.8 Flash
Model Studio (Intl)
Direct APIofficialofficial2026-08-28
QwenQwen3.7 Max (Promo)
Model Studio (Intl)
Direct APIofficialderived2026-07-20
QwenQwen3.7 Plus (Promo)
Model Studio (Intl)
Direct APIofficialderived2026-07-20
QwenQwen3.6 Plus
Model Studio (Intl)
Direct APIofficialderived2026-07-19
QwenQwen3.6 Flash
Model Studio (Intl)
Direct APIofficialderived2026-07-19
QwenQwen3.7 Flash
Model Studio (Intl)
Direct APIofficialderived2026-07-26
QwenQwen3.6 Max Preview
Model Studio (Intl)
Direct APIofficialderived2026-07-19
MistralMistral Medium 3.5Foundry ·Globalofficialno cache meter2026-07-23
MistralMistral Medium 3.5Foundry ·Data Zoneofficialno cache meter2026-07-23
MistralMistral Medium 3.5
Mistral direct API
Direct APIofficialofficial2026-07-23
MiniMaxMiniMax M2.5
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-07-26
MiniMaxMiniMax 3
Fireworks-hosted
Foundry ·Data Zoneofficialofficial2026-07-26
MiniMaxMiniMax M3
MiniMax direct API
Direct APIofficialofficial2026-07-29
MiniMaxMiniMax M2.7
MiniMax direct API
Direct APIofficialofficial2026-07-29
Upcoming

Scheduled rate changes

Every future boundary declared on a catalog lane’s published rate schedule, in order. A promotion or rate-change expiry fires once, on a calendar date; time-of-day pricing recurs on a schedule (e.g. a weekday peak/off-peak split).

  1. GLM-5.3-FlashGLM · Z.ai direct API
    9 Sep, 16:00 UTC
    Promotion / rate expiryList price (from September 10)
  2. 1 Jan 2027 UTC
    Promotion / rate expiryStandard (from 2027), Batch + 5 more
  3. 1 Jan 2027 UTC
    Promotion / rate expiryStandard (from 2027), Batch + 5 more
Coverage

Provider and source coverage

The official source registry every lane above traces back to, with each provider’s confidence breakdown across its tracked lanes.

ProviderLanesInput confidenceCached confidenceOfficial source
Claude87 official1 estimate7 official1 estimateAnthropic pricing
DeepSeek99 official8 official
1 without a cache meter
DeepSeek pricing
Gemini33 official3 officialGemini API pricing
GLM99 official9 officialZ.ai pricing
Grok55 official4 official
1 without a cache meter
xAI pricing
Kimi66 official6 officialMoonshot Kimi pricing
MiniMax44 official4 officialMiniMax pay-as-you-go
Mistral33 official1 official
2 without a cache meter
Mistral inference pricing
OpenAI / Azure OpenAI1313 official13 officialAzure Retail Prices API — Foundry Models
Qwen88 official2 official6 derivedModel Studio pricing