This Week in LLM Pricing — Jul 13, 2026
The inaugural entry in our weekly LLM pricing snapshot — exactly what the frontier models cost right now, pulled live from our own pricing API.
This is the first entry in our new weekly LLM pricing snapshot — here's exactly what the frontier models cost right now, pulled live from our own pricing API. Future entries will show week-over-week deltas once there's a prior snapshot to compare against.
By the LLMCalculator.net Research Team · Last updated: July 13, 2026
Frontier pricing snapshot (Jul 13, 2026)
USD per 1M tokens, list rate.
| Model | Provider | Input | Output | Context |
|---|---|---|---|---|
| GPT-5.6 Sol (preview) | OpenAI | $2.00 | $16.00 | 1M |
| GPT-5.6 Terra (preview) | OpenAI | $0.60 | $5.00 | 1M |
| GPT-5.6 Luna (preview) | OpenAI | $0.10 | $0.80 | 400K |
| GPT-5.5 (high) | OpenAI | $1.25 | $10.00 | 922K |
| Claude Fable 5 | Anthropic | $10.00 | $50.00 | 1M |
| Claude Sonnet 5 | Anthropic | $3.00 | $15.00 | 1M |
| Claude Opus 4.8 | Anthropic | $5.00 | $25.00 | 1M |
| Gemini 3.5 Pro | $1.50 | $12.00 | 2M | |
| Gemini 3.5 Flash | $0.30 | $2.50 | 1M | |
| Grok 4.3 (high) | xAI | $0.30 | $1.50 | 1M |
| DeepSeek V4 Pro (max) | DeepSeek | $0.27 | $1.10 | 1M |
| Qwen3.7 Max | Alibaba | $0.40 | $1.60 | 1M |
| MiMo V2.5 | Xiaomi | $0.015 | $0.18 | 1M |
Source: llmcalculator.net live pricing API (/api/v1/models), snapshot taken 2026-07-13.
Detailed breakdown
Across all 64 models we currently track, the cheapest is MiMo V2.5 from Xiaomi at $0.015 / $0.18 per 1M tokens, and the most expensive is Claude Fable 5 from Anthropic at $10 / $50 per 1M — a roughly 667× spread on input price alone. That gap keeps widening as ultra-cheap open-weights get productized and frontier reasoning models push in the other direction.
For most teams, the two default workhorse tiers to benchmark against first remain Claude Sonnet 5 at $3 / $15 and GPT-5.5 (high) at $1.25 / $10. They bracket the frontier price band where quality-per-dollar is best documented, and every other model on the list is worth evaluating as either a cost downgrade or a quality upgrade from one of those two.
About this series
This is the inaugural post in a new weekly series — going forward, each Monday we'll publish a fresh snapshot and, once there's a prior week to diff against, week-over-week deltas on every frontier model. Run the numbers on your own workload in the cost calculator, sanity-check against alternatives on the model comparison page, and see the full multi-provider view in the 2026 LLM price comparison or the cheapest LLM API 2026 roundup.