ByteDance UI-TARS 7B vs Meta Llama 3.2 1B Instruct
Side by side from live data: prices, limits, who serves each model and how reliably, and how they do on our chess arena.
In short
- Llama 3.2 1B Instruct is 1.8× cheaper at a typical mix of 3 input tokens per output token.
- UI-TARS 7B accepts longer prompts: 128K vs 60K tokens.
- UI-TARS 7B is the newer model (released 2025-07-22).
| UI-TARS 7B | Llama 3.2 1B Instruct | |
|---|---|---|
| Vendor | ByteDance | Meta |
| Released | 2025-07-22 | 2024-09-25 |
| Input / 1M | $0.10 | $0.027 |
| Output / 1M | $0.20 | $0.201 |
| Cached input / 1M | $0.10 | — |
| Blended / 1M (3:1) | $0.125 | $0.071 |
| Context | 128K | 60K |
| Max output | 2K | 54K |
| Accepts | text, image | text |
| Reasoning | No | No |
| Providers on OpenRouter | 1 | 1 |
| Cheapest provider, output | $0.20 | $0.201 |
| Best provider uptime, 24 h | 100% | 100% |
| Chess arena rating | not playing | not playing |
| Vendor API status | — | — |
Cost calculator
UI-TARS 7B
$9.00 per month$0.30 per day
Llama 3.2 1B Instruct
$4.63 per month$0.15 per day
Llama 3.2 1B Instruct saves $4.36 a month on this workload.
At list prices per 1M tokens; prompt caching and batch discounts are not included.
Compare UI-TARS 7B with
Compare Llama 3.2 1B Instruct with