Llama 3.2 3B Instruct vs Qwen-Omni Turbo.
model vs model · prices as of
Llama 3.2 3B Instruct takes 3 of 5 rounds against Qwen-Omni Turbo. It is also the cheaper of the two to run, at $0.08 per million tokens against $0.40.
Metawins 3 roundsLlama 3.2 3B Instruct
$0.08/M tokens
cheapest we can explain, at Inference · $0.0132 per 1,000 pages
131k context · open weights
Alibaba CloudQwen-Omni Turbo
$0.40/M tokens
cheapest we can explain, at Alibaba Cloud (China) · $0.0486 per 1,000 pages
33k context · tool calling · reads images
List price in (lower wins)
$0.05/M
$0.07/M
List price out (lower wins)
$0.33/M
$0.27/M
Cheapest we can explain (lower wins)
$0.08/M at Inference
$0.40/M at Alibaba Cloud (China)
Standard price (lower wins)
$0.48/M
$0.48/M
Cost per 1,000 pages (lower wins)
$0.0132
$0.0486
Context
131k tokens
33k tokens
Open weights
yes
no
Tool calling
no
yes
Structured output
yes
not stated
Reads images
no
yes
Reasoning
no
no
Released
25 Sept 2024
19 Jan 2025
Providers selling it
17
3
Price we measure against
OpenRouter
the maker's own
Trains on your prompts (no is better)
not stated
not stated
Providers selling both · 1
LLM Gateway$0.14 / $1.40
Llama 3.2 3B Instruct then Qwen-Omni Turbo, per million tokens
Go further
Add a third or fourth model, or switch the weighting, in the tool.
What moved in the last seven days
- Llama 3.2 3B Instruct
TrustedRouter$0.03 / $0.05 $0.05 / $0.35248% dearer ·
Frequently asked
- Which is cheaper, Llama 3.2 3B Instruct or Qwen-Omni Turbo?
- Llama 3.2 3B Instruct, at $0.08 per million tokens against $0.40 for Qwen-Omni Turbo, three parts input to one part output, read on 16 Sept 2026.
- Where is each one cheapest?
- Llama 3.2 3B Instruct: $0.08 at Inference. Qwen-Omni Turbo: $0.40 at Alibaba Cloud (China).
- Can I buy both from one provider?
- Yes. One provider sells both: LLM Gateway. The cheapest of them for the two together is LLM Gateway.
- Which holds more context?
- Llama 3.2 3B Instruct, at 131k tokens against 33k.
Prices exclude tax and are per million tokens, read 16 Sept 2026 from each provider's own catalog. The standard price blends three parts input to one part output; a page is 600 tokens in and 60 out. Bars are scaled to the larger value in each row.