What 19 leading language models cost via the API in October 2026 — input and output prices per 1M tokens, a blended price, context window and benchmark scores side by side. Updated monthly and free to reuse.
Click a column header to sort.
| Model | Provider | Input | Output | Blended | Context | Open weights | AA Index | Score |
|---|---|---|---|---|---|---|---|---|
| Claude Opus 5.5 | Anthropic | $4 | $20 | $8 | 1M | No | 58 | 99 |
| GPT-6 Astra | OpenAI | $10 | $50 | $20 | 1M | No | 53 | 98 |
| Claude Fable 5.1 | Anthropic | $10 | $50 | $20 | 1M | No | 53 | 98 |
| Claude Sonnet 5.5 | Anthropic | $2 | $10 | $4 | 1M | No | 56 | 97 |
| Fugu Ultra v2 | Sakana AI | $5 | $30 | $11.25 | 1M | No | — | 97 |
| Gemini 4 Argon (intro price) | $2 | $10 | $4 | 1M | No | 53 | 96 | |
| GPT-6.1 Sol | OpenAI | $2 | $10 | $4 | 1.05M | No | 52 | 96 |
| MiMo-V2.6-Pro | Xiaomi | $0.44 | $0.87 | $0.55 | 1M | Yes | 46.32 | 94 |
| Muse Wonder 1.0 | Meta | $4 | $20 | $8 | 1M | No | 52 | 94 |
| Grok 4.7 | xAI | $2 | $6 | $3 | 500K | No | 46 | 93 |
| Qwen 3.8-Max | Alibaba Cloud | $2 | $6 | $3 | 1M | Yes | 45 | 93 |
| DeepSeek V4.1 Flash | DeepSeek | $0.30 | $1.20 | $0.53 | 1M | Yes | 45 | 92 |
| GLM-5.3 | Zhipu AI | $1.40 | $4.40 | $2.15 | 1M | Yes | 45 | 92 |
| Kimi K3 | Moonshot AI | $3 | $15 | $6 | 1M | Yes | 44 | 92 |
| Claude Haiku 5.5 | Anthropic | $0.10 | $0.50 | $0.20 | 1M | No | 43 | 91 |
| MiniMax M3.1 Flash | MiniMax | $0.15 | $0.60 | $0.26 | 1M | Yes | 43 | 91 |
| Muse Spark Open (hosted API) | Meta | $1.25 | $4.25 | $2 | 1M | Yes | 39 | 88 |
| GPT-6 Luna | OpenAI | $0.10 | $0.50 | $0.20 | 1M | No | 38 | 87 |
| GPT-4o | OpenAI | $2.50 | $10 | $4.38 | 128K | No | — | 80 |
Blended = (3 × input + output) ÷ 4. AA Index = Artificial Analysis Intelligence Index where published. Score = AIVerso score (0–100).
The full table is available as a CSV file. You are welcome to use it in articles, newsletters and tools — please credit AIVerso with a link to this page.
Claude Haiku 5.5 (Anthropic) and GPT-6 Luna (OpenAI) are the cheapest at $0.10/$0.50 per 1M input/output tokens. Next come MiniMax M3.1 Flash at $0.15/$0.60 and DeepSeek V4.1 Flash at $0.30/$1.20.
Among models with an AIVerso score of 96 or higher, Claude Sonnet 5.5 is the cheapest at $2/$10 per 1M tokens. The most expensive is Claude Fable 5.1 at $10/$50.
Blended price = (3 × input price + output price) ÷ 4, assuming three input tokens for every output token, which is typical for chat and coding workloads. Your own mix may differ, so check both columns.
Standard pay-as-you-go list prices in US dollars per 1 million tokens, without batch or cache discounts. Where a provider has peak/off-peak rates, introductory prices or several tiers, we list the standard entry price and add a note.
At least once a month and whenever a major model launches or a provider changes its prices. Earlier monthly snapshots are kept so price changes can be tracked over time.
Yes. The table is free to reuse in articles and tools — please credit AIVerso with a link to this page. A CSV file with all prices is available for download.
Sources: list prices from each provider’s pricing page or API documentation, benchmark results from Artificial Analysis. Prices change often — check the provider before you commit. Last updated: October 23, 2026.