Updated May 24, 2026
llmleaderboard.in
Cheapest AI Models in 2026
Lowest published API pricing per million tokens — sorted by typical input cost. Output tokens often cost more; check both columns on the main leaderboard.
Nova Micro and Gemma 3 27B sit at the bottom of the price curve. DeepSeek V4 Flash offers a strong balance of low cost and solid GPQA/SWE scores for production workloads.
| # | Model | Provider | I/O cost | GPQA | SWE-Bench |
|---|---|---|---|---|---|
| 1 | Nova Micro | Amazon | $0.04 / $0.14 | — | — |
| 2 | Gemma 3 27B | $0.07 / $0.07 | 42.6% | 11.4% | |
| 3 | DeepSeek V4 Flash | DeepSeek | $0.08 / $0.28 | 79.4% | 68.2% |
| 4 | Gemini 2.0 Flash | $0.1 / $0.4 | 74.1% | 53.2% | |
| 5 | Gemini 2.5 Flash | $0.15 / $0.6 | 80.3% | 61.4% | |
| 6 | Mistral Small 4 | Mistral | $0.15 / $0.60 | 68.9% | 49.2% |
| 7 | GPT-5.4 Nano | OpenAI | $0.20 / $1.25 | 68.5% | 42.1% |
| 8 | Grok 4.1 Fast | xAI | $0.20 / $0.50 | 72.8% | 51.2% |
| 9 | Gemini 3.1 Flash-Lite | $0.25 / $1.50 | 86.9% | 62.8% | |
| 10 | DeepSeek V4 Pro | DeepSeek | $0.30 / $0.50 | 87.1% | 81% |
| 11 | MiniMax M2 | MiniMax | $0.30 / $1.20 | 78.2% | 68.5% |
| 12 | GPT-4.1 mini | OpenAI | $0.4 / $1.6 | 71.2% | 52.1% |
Cost vs quality
Cheapest is not always best. Budget models excel at high-volume simple tasks; use frontier models when accuracy on complex reasoning or coding matters. Compare head-to-head on the main compare tool.
See all 45 models with live benchmarks, speed, and pricing.
Open full LLM leaderboard →