Short answers to customer questions around the clock, on WhatsApp, a website or an app. Volume is high and every second of waiting counts.
8 models meet the requirements, ranked by monthly cost for this workload.
| Model | Arena score | Class | Context | Images | Input $/1M | Output $/1M | Monthly |
|---|---|---|---|---|---|---|---|
| Claude Haiku 5.5 Lowest cost | 1,425 | Fast | 1M | ✓ | $0.10 | $0.50 | $20 |
| GPT-6 Luna | 1,441 | Fast | 1.05M | ✓ | $0.10 | $0.50 | $20 |
| DeepSeek V4.1 Flash | 1,476 | Budget | 1M | ✓ | $0.15 | $0.60 | $25 |
| Llama 4 Maverick | 1,327 | Mid | 128K | ✓ | $0.19 | $0.65 | $34 |
| Qwen3.8-27B | 1,436 | Mid | 262K | ✓ | $0.15 | $1.88 | $57 |
| Gemini 3.5 Flash-Lite | 1,456 | Budget | 1.05M | ✓ | $0.30 | $2.50 | $81 |
| Gemini 3.8 Flash Promo | 1,496 | Workhorse | 1.05M | ✓ | $0.75 | $3.75 | $146 |
| GPT-6.1 Terra | — | Mid | 1.05M | ✓ | $1.00 | $5.00 | $190 |
— in a score column = not rated on that leaderboard yet.
Arena score: LMArena (text, style control), from human votes, published Oct 28, 2026, CC BY 4.0.
Our side-by-side tests for this task are in progress. Until they are published, run 10 to 20 of your own real examples on two or three candidates before you choose.
Independent leaderboards: LMArena · Artificial Analysis
How this list is built: models in the classes suited to the task that meet its hard requirements (image input, context size, open weights), ranked by price. A lower price does not mean better quality.