Cheapest Agent & Assistants APIs
Ranked by input price per 1M tokens, refreshed hourly. Free models first.
The cheapest Agent & Assistants API right now is LiquidAI: LFM2.5-1.2B-Thinking (free) (liquid): Free/1M input tokens.
Price Ranking
| # | Model | Vendor | Input /1M | Output /1M | Intelligence | Context | CPS | |
|---|---|---|---|---|---|---|---|---|
| 1 | LiquidAI: LFM2.5-1.2B-Thinking (free) | liquid | Free | Free | — | 33K | — | Details → |
| 2 | Arcee AI: Trinity Large Thinking (free) | arcee-ai | Free | Free | — | 262K | — | Details → |
| 3 | NVIDIA: Nemotron 3 Nano Omni (free) | nvidia | Free | Free | — | 256K | — | Details → |
| 4 | Baidu: ERNIE 4.5 21B A3B Thinking | baidu | $0.07 | $0.28 | — | 131K | — | Details → |
| 5 | Qwen: QwQ 32B | qwen | $0.15 | $0.58 | 10 | 131K | B | Details → |
| 6 | AllenAI: Olmo 3 32B Think | allenai | $0.15 | $0.50 | — | 66K | — | Details → |
| 7 | Qwen: Qwen3 Next 80B A3B Thinking | qwen | $0.15 | $1.20 | — | 262K | — | Details → |
| 8 | Qwen: Qwen3 VL 8B Thinking | qwen | $0.18 | $2.10 | — | 131K | — | Details → |
| 9 | Qwen: Qwen3 30B A3B Thinking 2507 | qwen | $0.20 | $2.40 | — | 82K | — | Details → |
| 10 | Arcee AI: Trinity Large Thinking | arcee-ai | $0.25 | $0.80 | — | 262K | — | Details → |
| 11 | DeepSeek: R1 Distill Qwen 32B | deepseek | $0.29 | $0.29 | 8 | 128K | B | Details → |
| 12 | Qwen: Qwen3 VL 30B A3B Thinking | qwen | $0.29 | $1.00 | — | 262K | — | Details → |
FAQ
What is the cheapest Agent & Assistants model?
By input price, it is LiquidAI: LFM2.5-1.2B-Thinking (free) (Free/1M tokens, liquid). Full ranking above, refreshed hourly.
Are there free models?
Yes. Models marked "Free" have $0 input pricing, usually rate-limited — good for dev/test. For production, pick S/A CPS-grade paid models.
How do I choose a Agent & Assistants model?
Shortlist by price here, then check intelligence/coding index and context fit, and use head-to-head comparisons for the final call.