- Rankings
- Cheapest AI models by output token price
Cheapest AI models by output token price
Ranked on output price alone. This is the table to use for reasoning models and agents, where thinking tokens bill as output and routinely outnumber the visible answer several times over.
List prices checked
All 25, ranked
| # | Name | Provider | Output | Notes |
|---|---|---|---|---|
| 1 | DeepSeek-V4-Flash | DeepSeek | $0.280/1M tokens | 1M context · $0.140 in / $0.280 out |
| 2 | Mistral Small | Mistral AI | $0.300/1M tokens | 131K context · $0.100 in / $0.300 out |
| 3 | GPT-5 nano | OpenAI | $0.400/1M tokens | 400K context · $0.050 in / $0.400 out |
| 4 | Gemini 2.5 Flash-Lite | $0.400/1M tokens | 1M context · $0.100 in / $0.400 out | |
| 5 | Grok 3 mini | xAI | $0.500/1M tokens | 131K context · $0.300 in / $0.500 out |
| 6 | GPT-4o mini | OpenAI | $0.600/1M tokens | 128K context · $0.150 in / $0.600 out |
| 7 | Cohere Command R | Cohere | $0.600/1M tokens | 128K context · $0.150 in / $0.600 out |
| 8 | DeepSeek-V4-Pro | DeepSeek | $0.870/1M tokens | 1M context · $0.435 in / $0.870 out |
| 9 | Codestral | Mistral AI | $0.900/1M tokens | 262K context · $0.300 in / $0.900 out |
| 10 | GPT-4.1 mini | OpenAI | $1.60/1M tokens | 1M context · $0.400 in / $1.60 out |
| 11 | GPT-5 mini | OpenAI | $2.00/1M tokens | 400K context · $0.250 in / $2.00 out |
| 12 | Gemini 2.5 Flash | $2.50/1M tokens | 1M context · $0.300 in / $2.50 out | |
| 13 | OpenAI o4-mini | OpenAI | $4.40/1M tokens | 200K context · $1.10 in / $4.40 out |
| 14 | Claude Haiku 4.5 | Anthropic | $5.00/1M tokens | 200K context · $1.00 in / $5.00 out |
| 15 | Mistral Large | Mistral AI | $6.00/1M tokens | 131K context · $2.00 in / $6.00 out |
| 16 | GPT-4.1 | OpenAI | $8.00/1M tokens | 1M context · $2.00 in / $8.00 out |
| 17 | OpenAI o3 | OpenAI | $8.00/1M tokens | 200K context · $2.00 in / $8.00 out |
| 18 | GPT-5 | OpenAI | $10.00/1M tokens | 400K context · $1.25 in / $10.00 out |
| 19 | GPT-4o | OpenAI | $10.00/1M tokens | 128K context · $2.50 in / $10.00 out |
| 20 | Gemini 2.5 Pro | $10.00/1M tokens | 1M context · $1.25 in / $10.00 out | |
| 21 | Cohere Command A | Cohere | $10.00/1M tokens | 256K context · $2.50 in / $10.00 out |
| 22 | Claude Sonnet 5 | Anthropic | $15.00/1M tokens | 200K context · $3.00 in / $15.00 out |
| 23 | Grok 4 | xAI | $15.00/1M tokens | 256K context · $3.00 in / $15.00 out |
| 24 | Claude Opus 5 | Anthropic | $25.00/1M tokens | 200K context · $5.00 in / $25.00 out |
| 25 | Claude Fable 5 | Anthropic | $50.00/1M tokens | 200K context · $10.00 in / $50.00 out |
Source: DeepSeek (August 2026) · Mistral AI (May 2026) · OpenAI (May 2026) · Google (May 2026) · xAI (May 2026) · Cohere (May 2026). Published list prices only — committed-use discounts, regional variation, and per-request charges are not included. Not yet independently re-checked against the vendor’s page.
Questions
- What is the cheapest ai models by output token price in 2026?
- DeepSeek-V4-Flash from DeepSeek tops this ranking at $0.280/1M tokens. The full table below shows all 25 entries with the same measure applied to each.
- How is this ranking calculated?
- Ranked on output price alone. This is the table to use for reasoning models and agents, where thinking tokens bill as output and routinely outnumber the visible answer several times over. Every figure is the provider's own published list price, linked in the source column and stamped with the month it was checked.
- Why is the cheapest option not always the right one?
- Because a single axis never captures the whole bill. A provider with the lowest storage rate may have the highest egress rate; a model with the cheapest input tokens may be the most expensive for agentic work. The comparison pages model both sides against identical workloads, which is the check worth doing before switching.
Other rankings
- Cheapest CDN providers by egress price
- Cheapest object storage per GB-month
- Cheapest VPS and cloud servers
- Cheapest cloud egress rates, all categories
- Cloud services with no egress fees
- Cheapest managed database hosting
- Cheapest application hosting platforms
- Cloud services with the best free tiers
- Cheapest AI model APIs by blended token price
- Cheapest frontier AI models
- AI models with the longest context windows