fix(ms-ai-architect): re-baseline kostnadslaget (GPT-5 $1.25/$10, én prissannhet, 10 filer)

Audit-funn Tema B + action-items 149/150. Verifiseringen avdekket at kostnadslag-råten var bredere enn auditens 3 navngitte filer (10 filer totalt).

- GPT-5 flagship $10-15/$40-60 -> $1.25/$10.00 (verifisert juni 2026, 8-12x for hoy). gpt-5-mini $0.25/$2.00, gpt-5-nano $0.05/$0.40. Fjernet falske "Verified MCP"-stempel pa estimat-band.

- deterministic-cost-calculation-model.md gjort til ENESTE kanoniske prissannhet (§1.1). Strippet duplikat pris-tabell fra cost-models.md (rotarsak til GPT-4o $10/$30-motsetning) -> peker til registeret.

- Reconcile-funn utover audit: deterministic hadde GPT-4.1-serien HALVERT ($1/$4 og $0.20/$0.80) -> verifisert $2.00/$8.00 og $0.40/$1.60. multi-model-strategy-costs.md NOK-tabell var fabrikkert 10-47x for hoy (gpt-4o 250/1000 NOK -> 26/105). model-selection (GPT-5 $3/$12), reasoning-models (GPT-5 $20-30/$80-100), token-counting (gpt-4.1 cached $1->$0.50), small-language-models (GPT-4o output 60->105 NOK) rettet.

- Recalk alle NOK-scenarioer reproduserbart: gpt-5 ~= gpt-4.1 i kost (ikke 30-55x dyrere som gammel tekst pasto) - billigere input ($1.25 vs $2.00), dyrere output. Reell kostnadsdriver = resonnerings-tokens (fakturert som output).

- Fjernet GPT-3.5 Turbo-anbefaling (utfaset). Alignet kurs 11->10.5 NOK/USD i cost-kommando. Cost-agent eksempel: legacy GPT-4 -> GPT-4o-mini, fjernet falskt Verified-stempel.

- kb-update apply cost-bucket deferret til task #8 (kommando-feature, ikke tall-justering).

Verifisert mot OpenAI API pricing + Azure OpenAI pricing via WebSearch 2026-06-18. validate-plugin.sh 219 PASS.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Claude-Session: https://claude.ai/code/session_01REiKFhP4w6xGXXqWKpPCJJ
This commit is contained in:
Kjell Tore Guttormsen 2026-06-18 12:45:21 +02:00
commit 84e3cfaa03
10 changed files with 143 additions and 112 deletions

View file

@ -1,8 +1,9 @@
# Deterministisk kostnadsberegningsmodell for AI-arkitekturvurderinger
**Sist oppdatert:** 2026-05 (v1.1)
**Sist oppdatert:** 2026-06 (v1.2)
**Status:** GA
**Category:** Cost Optimization & FinOps for AI
**Rolle:** Kanonisk prissannhet — eneste autoritative enhetspris-register for pluginen
---
@ -29,16 +30,27 @@ Modellen sikrer at:
| Modell | Input (per 1M tokens) | Cached Input (per 1M tokens) | Output (per 1M tokens) | Kilde | Verifisert |
|--------|----------------------|------------------------------|------------------------|-------|------------|
| **GPT-4o** | $2.50 | $1.25 | $10.00 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-02 |
| **GPT-4o-mini** | $0.15 | $0.075 | $0.60 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-02 |
| **GPT-5** | $1.25 | $0.125 | $10.00 | [OpenAI API pricing](https://developers.openai.com/api/docs/pricing/) / WebSearch | 2026-06 |
| **GPT-5-mini** | $0.25 | — | $2.00 | [OpenAI API pricing](https://developers.openai.com/api/docs/pricing/) | 2026-06 |
| **GPT-5-nano** | $0.05 | — | $0.40 | [OpenAI API pricing](https://developers.openai.com/api/docs/pricing/) / WebSearch | 2026-06 |
| **GPT-4o** | $2.50 | $1.25 | $10.00 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-06 |
| **GPT-4o-mini** | $0.15 | $0.075 | $0.60 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-06 |
| **o3-mini** | $1.10 | $0.55 | $4.40 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-02 |
| **GPT-4.1** | $1.00 | — | $4.00 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-02 |
| **GPT-4.1-mini** | $0.20 | — | $0.80 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-02 |
| **GPT-4.1** | $2.00 | $0.50 | $8.00 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-06 |
| **GPT-4.1-mini** | $0.40 | $0.10 | $1.60 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) | 2026-06 |
| **GPT-4.1-nano** | $0.10 | $0.025 | $0.40 | [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) (interpolert) | 2026-06 |
| **text-embedding-3-small** | $0.02 | — | — | [OpenAI Pricing](https://developers.openai.com/api/docs/pricing/) | 2026-02 |
| **text-embedding-3-large** | $0.13 | — | — | [OpenAI Pricing](https://developers.openai.com/api/docs/pricing/) | 2026-02 |
**Denne tabellen er den kanoniske prissannheten for pluginen.** Alle andre KB-filer (inkl. `cost-models.md` og `gpt5-gpt41-pricing-models.md`) skal referere hit for rå token-priser, ikke duplisere egne tabeller. Ved konflikt vinner denne.
**Merknad:** Azure OpenAI-priser er typisk identiske med OpenAI API-priser for Global Standard deployment. Regional deployment og Data Zone deployment kan ha andre priser. Priser over er per 1 million tokens (1M), ikke per 1K.
**Merknader til enkeltrader (verifiseringsplikt):**
- **GPT-5-mini:** $0.25/$2.00 er OpenAIs publiserte listepris (lansering aug 2025). Enkelte pris-aggregatorer rapporterer en redusert rate ($0.125/$1.00) per 2026 — verifiser mot Azure prisside før formelt estimat. Det konservative tallet ($0.25/$2.00) brukes her for ikke å under-budsjettere.
- **GPT-4.1-serien:** Korrigert 2026-06 fra tidligere halverte verdier ($1.00/$4.00 og $0.20/$0.80) til verifisert listepris ($2.00/$8.00 og $0.40/$1.60). Kilde: Azure Content Understanding-priseksempler + WebSearch juni 2026.
- **Nyere GPT-5-generasjoner** (gpt-5.2 $1.75/$14, gpt-5.4 $2.50/$15, gpt-5.5 $5.00/$30) er verifisert juni 2026 men håndteres i modellkatalogen (`platforms/model-catalog-2026.md`), ikke her. Bruk dem kun når et use case eksplisitt krever den generasjonen.
### 1.2 Azure AI Search — Månedlig per Search Unit (SU)
| Tier | Pris per SU/måned (USD) | Lagring per partisjon | Maks SU | Kilde | Verifisert |