PRICING WATCH
AI Model Pricing Watch
What frontier models actually cost, in one table, with a date on it. AI answers will quote you last quarter's prices with total confidence; this page tells you when a human last checked. If the date looks stale, assume the numbers are too — that's the point of printing it.
Prices last verified: July 19, 2026
OpenAI
| Model | Input /1M | Output /1M | Context | Tier |
|---|
| GPT-5.6 Sol | $5.00 | $30.00 | 1M | Frontier |
| GPT-5.6 Terra | $2.50 | $15.00 | 1M | Balanced |
| GPT-5.6 Luna | $1.00 | $6.00 | 1M | Budget |
| GPT-5.5 | $5.00 | $30.00 | 200K | Frontier |
| GPT-5.4 | $2.50 | $15.00 | 128K | Balanced |
| GPT-5.4 Mini | $0.75 | $4.50 | 128K | Budget |
| GPT-4.1 | $2.00 | $8.00 | 1M | Balanced |
| GPT-4.1 Nano | $0.10 | $0.40 | 1M | Budget |
Meta
| Model | Input /1M | Output /1M | Context | Tier |
|---|
| Muse Spark 1.1 | $1.25 | $4.25 | 1M | Balanced |
Anthropic
| Model | Input /1M | Output /1M | Context | Tier |
|---|
| Claude Opus 4.7 | $5.00 | $25.00 | 1M | Frontier |
| Claude Sonnet 4.6 | $3.00 | $15.00 | 200K | Balanced |
| Claude Haiku 4.5 | $1.00 | $5.00 | 200K | Budget |
Google
| Model | Input /1M | Output /1M | Context | Tier |
|---|
| Gemini 3.1 Pro | $2.00 | $12.00 | 1M | Frontier |
| Gemini 3 Flash | $0.50 | $3.00 | 1M | Budget |
DeepSeek
| Model | Input /1M | Output /1M | Context | Tier |
|---|
| DeepSeek V3 (chat) | $0.27 | $1.10 | 64K | Budget |
Mistral
| Model | Input /1M | Output /1M | Context | Tier |
|---|
| Mistral Large 3 | $0.50 | $1.50 | 256K | Balanced |
All prices in USD per 1 million tokens, standard API rates — no volume discounts or cache pricing.
What changed — July 2026
- OpenAI launched the GPT-5.6 family (July 9): Sol at $5/$30, Terra at $2.50/$15, Luna at $1/$6 — all with a 1M-token context window. The 5.5/5.4 line remains available for now.
- Meta entered the paid-API business: Muse Spark 1.1 at $1.25/$4.25 with 1M context, US-only public preview with $20 in starter credits.
- Google's Gemini 3.5 Pro is still unpriced — the launch slipped over coding performance, so 3.1 Pro remains the priced flagship.
Prices are inputs. What you actually pay depends on your workload.