Zum Inhalt springen
optera

Änderungsprotokoll

KI-Preis- Aktualisierungen

Neue Modelle, Preisänderungen und Einstellungen – jeweils mit Link zur Seite des Anbieters.
Neues Modell

Claude Haiku 5.5 launches at a tenth of Haiku 4.5's price

Anthropic released Claude Haiku 5.5 at $0.10 / $0.50 per million tokens for prompts up to 100K tokens and $0.50 / $2.50 above, with a 1M-token context. Anthropic's launch table puts it ahead of GPT-6 Luna on every benchmark both were run on. Its tokenizer is the Claude 4.7+ one, which turns the same text into about 30% more tokens than Haiku 4.5's.

Feature

Every price re-checked against the providers — and retired models marked

All 44 models were checked against each provider's own pricing and deprecation pages. Twelve current models were added, thirteen retired ones are now labelled with the date they were shut down and what replaced them, and every cost estimate now accounts for tokenizer differences — Anthropic documents that Claude 4.7 and later produce about 30% more tokens for the same text.

Neues Modell

GPT-6.1 Sol replaces GPT-6 Sol at the same price

OpenAI lists GPT-6.1 Sol at $2 / $10 per million tokens — the same as GPT-6 Sol a week earlier — with cached input halved to $0.10. Prompts over 272K tokens cost $4 / $15.

Neues Modell

Claude Sonnet 5.5 launches at Sonnet 5's price

Anthropic released Claude Sonnet 5.5 at $2 / $10 per million tokens, unchanged from Sonnet 5, with a 1M-token context. Like every Claude since 4.7, its tokenizer turns the same text into about 30% more tokens than earlier Claude — worth counting against GPT-6.1 Sol at the same list price.

Neues Modell

Claude Opus 5.5 costs 20% less than Opus 5

Anthropic's new Opus lists at $4 / $20 per million tokens, down from $5 / $25. Cache reads cost $0.20 per million (5% of the input price), and fast mode is $8 / $40.

Neues Modell

GPT-6 Sol and GPT-6 Luna

OpenAI added GPT-6 Sol at $2 / $10 and GPT-6 Luna at $0.10 / $0.50 per million tokens, both with a 1.05M-token context. Luna lists at half the price of GPT-5.6 Luna.

Neues Modell

Grok 4.7 keeps Grok 4.5's price

xAI's Grok 4.7 costs $2 / $6 per million tokens for prompts under 200K tokens and $4 / $12 above, with a 500K-token context — the same as Grok 4.5.

Neues Modell

DeepSeek V4.1 Flash

DeepSeek's deepseek-flash endpoint now serves V4.1 Flash at $0.30 / $1.20 per million tokens in peak hours and half that off-peak, with a 1M-token context.

Neues Modell

GPT-6 Astra, OpenAI's new flagship

$10 / $50 per million tokens for prompts up to 272K tokens and $20 / $75 above, with a 1.05M-token context — about twice GPT-5.6 Sol's regular $5 / $30.

Neues Modell

Gemini 3.8 Flash, on an introductory price

Google lists Gemini 3.8 Flash at $0.75 / $3.75 per million tokens until 31 December 2026, rising to $1.50 / $7.50 from 1 January 2027. 1M-token context.

Neues Modell

Claude Fable 5.1: same price, much cheaper cache reads

Fable 5.1 keeps Fable 5's $10 / $50 per million tokens, but cache reads drop to $0.25 per million — 2.5% of the input price, against the usual 10%.

Preissenkung

GPT-5.6 Sol on a promotional $4 / $20

OpenAI cut GPT-5.6 Sol from $5 / $30 to $4 / $20 per million tokens, guaranteed until at least 21 November 2026.

Preissenkung

GPT-5.6 Luna 80% cheaper, Terra 20% cheaper

Luna fell from $1 / $6 to $0.20 / $1.20 per million tokens and Terra from $2.50 / $15 to $2 / $12.

Neues Modell

Claude Opus 5 at half of Fable's price

Anthropic released Claude Opus 5 at $5 / $25 per million tokens with a 1M-token context — half the price of Claude Fable 5.

Eingestellt

DeepSeek retires its V3-era API names

The deepseek-chat and deepseek-reasoner names, which served DeepSeek V3.2, stopped accepting requests. DeepSeek's API now lists only V4 Pro and the Flash line.

Neues Modell

Grok 4.5

xAI's Grok 4.5 costs $2 / $6 per million tokens for prompts under 200K tokens, with a 500K-token context.

Neues Modell

Claude Sonnet 5 at $2 / $10

Anthropic launched Claude Sonnet 5 at an introductory $2 / $10 per million tokens, then made it the standard price and cancelled a planned rise to $3 / $15.

Feature

AI model leaderboard

Current models ranked in six categories — value, speed, coding, reasoning, creative and vision — on iOPTERA's editorial scores, with links to each model's page.

Feature

Savings calculator

Pick the model you run and a model you are considering, set your volume, and see the monthly and yearly difference. The URL carries your inputs, so a colleague can check the maths.

Feature

Stack optimizer: one model or several?

Select up to four kinds of work and compare a dedicated model per task against a single model for everything.

Neues Modell

Claude Fable 5, a tier above Opus

Anthropic introduced Claude Fable 5 at $10 / $50 per million tokens with a 1M-token context, priced above its Opus line.

Eingestellt

Gemini 2.0 Flash shut down

Google shut down Gemini 2.0 Flash and Gemini 2.0 Flash-Lite on its API. Requests to them now fail.

Neues Modell

Claude Opus 4.8

Anthropic released Claude Opus 4.8 at $5 / $25 per million tokens with a 1M-token context.

Eingestellt

Claude 3 Haiku retired

Anthropic retired Claude 3 Haiku on its API. Anthropic's recommended replacement is Claude Haiku 4.5.

Eingestellt

Claude 3 Opus retired

Anthropic retired Claude 3 Opus on its API. It had cost $15 / $75 per million tokens; Claude Opus 5.5 now lists at $4 / $20.