Phi 4 vs Qwen Turbo.
model vs model · prices as of
Qwen Turbo takes 3 of 5 rounds against Phi 4. It is also the cheaper of the two to run, at $0.22 per million tokens against $0.35.
Phi 4
$0.35/M tokens
OpenRouter list price, nothing below it is accounted for · $0.0504 per 1,000 pages
16k context · open weights
Alibaba Cloudwins 3 roundsQwen Turbo
$0.22/M tokens
cheapest we can explain, at Alibaba Cloud (China) · $0.0316 per 1,000 pages
1M context · tool calling
List price in (lower wins)
$0.07/M
$0.05/M
List price out (lower wins)
$0.14/M
$0.20/M
Cheapest we can explain (lower wins)
none below list
$0.22/M at Alibaba Cloud (China)
Standard price (lower wins)
$0.35/M
$0.35/M
Cost per 1,000 pages (lower wins)
$0.0504
$0.0316
Context
16k tokens
1M tokens
Open weights
yes
no
Tool calling
no
yes
Structured output
yes
not stated
Reads images
no
no
Reasoning
no
yes
Released
10 Jan 2025
1 Nov 2024
Providers selling it
9
7
Price we measure against
OpenRouter
the maker's own
Trains on your prompts (no is better)
no provider to check
not stated
Providers selling both · 2
Eden AI$0.35 / $0.23
Requesty$0.35 / $0.35
Phi 4 then Qwen 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
- Qwen Turbo
Ofox$0.05 / $0.09 $0.04 / $0.099% cheaper ·
Frequently asked
- Which is cheaper, Phi 4 or Qwen Turbo?
- Qwen Turbo, at $0.22 per million tokens against $0.35 for Phi 4, three parts input to one part output, read on 16 Sept 2026.
- Where is each one cheapest?
- Phi 4: nobody goes below OpenRouter's own price with a reason we can point at, so $0.35 is the number to beat. Qwen Turbo: $0.22 at Alibaba Cloud (China).
- Can I buy both from one provider?
- Yes. 2 providers sell both: Eden AI, Requesty. The cheapest of them for the two together is Eden AI.
- Which holds more context?
- Qwen Turbo, at 1M tokens against 16k.
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.