The reports support procurement, IT, privacy and management decisions. Prices, regions, contract terms and privacy requirements must be reviewed individually before procurement.
How to read the figures
Price rows
A model may have several rows, such as input, output, cache read or an additional pricing route. This is not identical to the number of models.
Quality score
The score reflects how well a price entry is supported by sources, primary-source evidence, parsers, plausibility checks and anomaly checks.
Review required
These entries remain visible but should be used carefully because of source, aggregator route, region, model alias or price-component uncertainty.
Supported additional prices
Additional official or plausibly supported price routes that improve coverage while remaining transparently marked.
Cache Read / Cache Write
Some providers charge cheaper cache reads or separate cache writes. These are not directly identical to standard input and output prices.
Provider overview
The table summarizes providers by public price coverage and quality. It is orientation, not a complete procurement recommendation.
| Provider | Price rows | Models | Avg. quality | Status | Review rows | Primary source seen | Official price aligned |
|---|---|---|---|---|---|---|---|
| OpenAI | 279 | 104 | 29.8 | prüfen | 279 | 54 | 26 |
| 193 | 67 | 28.6 | prüfen | 193 | 85 | 14 | |
| Alibaba / Qwen | 190 | 109 | 33.1 | prüfen | 190 | 95 | 45 |
| OpenRouter | 140 | 55 | 11.6 | prüfen | 140 | 0 | 0 |
| Anthropic | 138 | 42 | 43.0 | prüfen | 138 | 68 | 44 |
| Meta / Llama | 71 | 50 | 22.8 | prüfen | 71 | 0 | 0 |
| DeepInfra | 69 | 69 | 19.0 | prüfen | 69 | 0 | 0 |
| Mistral AI | 65 | 33 | 26.8 | prüfen | 65 | 6 | 0 |
| DeepSeek | 59 | 33 | 29.8 | prüfen | 59 | 15 | 9 |
| Moonshot / Kimi | 42 | 20 | 23.8 | prüfen | 42 | 0 | 0 |
| Z.ai / GLM | 39 | 13 | 27.0 | prüfen | 39 | 0 | 0 |
| NVIDIA | 30 | 13 | 7.7 | prüfen | 30 | 0 | 0 |
| xAI | 27 | 13 | 31.7 | prüfen | 27 | 6 | 3 |
| MiniMax | 24 | 9 | 27.0 | prüfen | 24 | 0 | 0 |
| Together AI | 13 | 5 | 19.0 | prüfen | 13 | 0 | 0 |
| ByteDance | 12 | 6 | 13.0 | prüfen | 12 | 0 | 0 |
| Amazon / AWS | 11 | 5 | 32.8 | prüfen | 11 | 2 | 0 |
| Perplexity | 11 | 5 | 27.0 | prüfen | 11 | 0 | 0 |
| Cohere | 10 | 5 | 25.0 | prüfen | 10 | 0 | 0 |
| Nous Research | 8 | 4 | 13.0 | prüfen | 8 | 0 | 0 |
| Tencent | 8 | 3 | 27.0 | prüfen | 8 | 0 | 0 |
| Arcee AI | 5 | 2 | 13.0 | prüfen | 5 | 0 | 0 |
| Microsoft / Azure | 4 | 2 | 27.0 | prüfen | 4 | 0 | 0 |
| AI21 | 2 | 1 | 27.0 | prüfen | 2 | 0 | 0 |
| AllenAI | 2 | 1 | 13.0 | prüfen | 2 | 0 | 0 |
| Baidu | 2 | 1 | 27.0 | prüfen | 2 | 0 | 0 |
| Liquid AI | 2 | 1 | 3.0 | prüfen | 2 | 0 | 0 |
Models with higher data quality
| Provider | Model | Price route | Component | EUR / 1M | Score | Quality | Source |
|---|---|---|---|---|---|---|---|
| DeepSeek | DeepSeek: DeepSeek V4 Pro 0813 | OpenRouter Models API | cache_read | €0.00 | 84.0 | high | Source |
| Anthropic | Anthropic Claude Haiku Latest | OpenRouter Models API | cache_read | €0.09 | 84.0 | high | Source |
| Anthropic | Anthropic: Claude Haiku 4.5 | OpenRouter Models API | cache_read | €0.09 | 84.0 | high | Source |
| DeepSeek | DeepSeek: DeepSeek V4 Flash 0423 | OpenRouter Models API | input | €0.12 | 84.0 | high | Source |
| DeepSeek | DeepSeek: DeepSeek V4 Flash 0731 | OpenRouter Models API | input | €0.12 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3.5-35B-A3B | OpenRouter Models API | input | €0.19 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3 235B A22B Thinking 2507 | OpenRouter Models API | input | €0.20 | 84.0 | high | Source |
| Anthropic | Anthropic: Claude Sonnet 4.5 | OpenRouter Models API | cache_read | €0.26 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3 Coder 480B A35B | OpenRouter Models API | input | €0.26 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3.7 Plus | OpenRouter Models API | cache_write | €0.35 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3.6 Plus | OpenRouter Models API | cache_write | €0.35 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3 235B A22B | OpenRouter Models API | input | €0.39 | 84.0 | high | Source |
| OpenAI | OpenAI: GPT-3.5 Turbo | OpenRouter Models API | input | €0.43 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3.6 27B | OpenRouter Models API | input | €0.52 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3 Coder Plus | OpenRouter Models API | cache_write | €0.70 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3 Coder Flash | OpenRouter Models API | output | €0.84 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3.6 Flash | OpenRouter Models API | output | €0.97 | 84.0 | high | Source |
| OpenAI | OpenAI: GPT-5.3-Codex | OpenRouter Models API | input | €1.51 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3.5 397B A17B | OpenRouter Models API | output | €2.02 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3 VL 30B A3B Thinking | OpenRouter Models API | output | €2.07 | 84.0 | high | Source |
| Alibaba / Qwen | Qwen: Qwen3.5-122B-A10B | OpenRouter Models API | output | €2.07 | 84.0 | high | Source |
| Anthropic | Anthropic: Claude Sonnet 4 | OpenRouter Models API | input | €2.59 | 84.0 | high | Source |
| Anthropic | Anthropic: Claude Sonnet 4.6 | OpenRouter Models API | input | €2.59 | 84.0 | high | Source |
| Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview) | OpenRouter Models API | output | €2.59 | 84.0 | high | Source | |
| Google: Nano Banana 2 (Gemini 3.1 Flash Image) | OpenRouter Models API | output | €2.59 | 84.0 | high | Source | |
| OpenAI | OpenAI: GPT-5.5 | OpenRouter Models API | input | €4.32 | 84.0 | high | Source |
| OpenAI | OpenAI: GPT-5.6 Sol | OpenRouter Models API | input | €4.32 | 84.0 | high | Source |
| Google: Nano Banana Pro (Gemini 3 Pro Image Preview) | OpenRouter Models API | output | €10.37 | 84.0 | high | Source | |
| Google: Nano Banana Pro (Gemini 3 Pro Image) | OpenRouter Models API | output | €10.37 | 84.0 | high | Source | |
| Anthropic | Anthropic: Claude Opus 4.5 | OpenRouter Models API | output | €21.61 | 72.0 | medium | Source |
Further analyses
Detailed CSV exports, API data, model comparison reports, procurement lists and custom scenario analyses can be provided to customers and project inquiries separately.