BroadRouter

Qwen2.5 72B vs Qwen 2.5 7B API pricing

Updated

Qwen2.5 72B: USD 0.36 input, USD 0.40 output per 1M tokens (DeepInfra). Qwen 2.5 7B: USD 0.05 input, USD 0.05 output per 1M tokens (SiliconFlow). At 3 input tokens per output token, Qwen2.5 72B costs 7.4× as much as Qwen 2.5 7B (USD 0.37 vs USD 0.05 per 1M).

Qwen2.5 72B and Qwen 2.5 7B are neighbouring sizes of the same Qwen generation (72B and 7B parameters, by name). Price only: every number is computed from prices read from each provider's own pricing page (linked per cell, newest check 2026-11-15), with no claim about quality, speed or accuracy.

How do Qwen2.5 72B and Qwen 2.5 7B differ on price?

Qwen2.5 72B and Qwen 2.5 7B prices side by side

USD per 1M tokens, standard tier. "Cheapest" is the lowest price at 3 input tokens to 1 output token; each line can come from a different provider.

Qwen2.5 72B and Qwen 2.5 7B API prices compared
PriceQwen2.5 72BQwen 2.5 7B
Model pageQwen2.5-72B-Instruct
Qwen · 3 providers
Qwen 2.5 7B
Qwen · 3 providers
Cheapest standard price, blended 3:1USD 0.37
USD 0.36 in / USD 0.40 out at DeepInfra
DeepInfra page checked 2026-11-14
USD 0.05 lower
USD 0.05 in / USD 0.05 out at SiliconFlow
SiliconFlow page checked 2026-11-14
Cheapest inputUSD 0.36
DeepInfra source
USD 0.05 lower
SiliconFlow source
Cheapest cached inputNo standard price lists a cache price; cached tokens bill at the input price.USD 0.136
Featherless source
Cheapest outputUSD 0.40
DeepInfra source
USD 0.05 lower
SiliconFlow source
Maker's own APIQwen publishes no paid standard price on its own API that we found.Qwen publishes no paid standard price on its own API that we found.
Median standard priceUSD 0.37 / USD 0.40
input / output, 3 prices
USD 0.05 / USD 0.13
input / output, 3 prices
Context window32,768 tokens, largest stated32,768 tokens, largest stated; from 32,000 at some providers
Providers with a standard price33

Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).

What do Qwen2.5 72B and Qwen 2.5 7B cost for a real workload?

30 days at each model's cheapest standard price for the workload (presets of the cost calculator; cached input at the cached price where listed).

Monthly cost of three workloads on Qwen2.5 72B and Qwen 2.5 7B
WorkloadQwen2.5 72BQwen 2.5 7BDifference
Support chatbot
2,000 requests a day, 1,500 input tokens (60% cached), 300 output
USD 39.60
DeepInfra
USD 5.40
SiliconFlow
USD 34.20
Qwen 2.5 7B lower
RAG over documents
1,000 requests a day, 8,000 input tokens (30% cached), 500 output
USD 92.40
DeepInfra
USD 12.75
SiliconFlow
USD 79.65
Qwen 2.5 7B lower
Coding agent (one developer)
1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning
exceeds contextexceeds context—

Qwen 2.5 7B is cheaper on 2 workloads. The widest gap is on Support chatbot: the dearer model costs 7.3× as much as the cheaper one.

Does the cheaper model change with your input:output ratio?

Qwen 2.5 7B is cheaper at every mix of input and output tokens, before cache discounts: its cheapest input price and its cheapest output price are both at or below Qwen2.5 72B's, so the input:output ratio never changes the answer.

Price per 1M tokens of Qwen2.5 72B and Qwen 2.5 7B at different input:output mixes
Input : output tokensQwen2.5 72BQwen 2.5 7B
1 : 3 (long answers)USD 0.39
DeepInfra
USD 0.05 lower
SiliconFlow
1 : 1USD 0.38
DeepInfra
USD 0.05 lower
SiliconFlow
3 : 1 (the site’s blend)USD 0.37
DeepInfra
USD 0.05 lower
SiliconFlow
10 : 1 (retrieval, long prompts)USD 0.3636
DeepInfra
USD 0.05 lower
SiliconFlow
30 : 1 (agents, big contexts)USD 0.3613
DeepInfra
USD 0.05 lower
SiliconFlow

USD per 1M tokens at the cheapest standard provider for each mix, before cache discounts. Thinking tokens bill as output.

What do Qwen2.5 72B and Qwen 2.5 7B cost per character in other languages?

Not measured: the real cost by language dataset covers only tokenizers we could measure from published files, and Qwen2.5 72B and Qwen 2.5 7B are not among them. Count tokens on your own text with the counters below.

Check it on your own traffic

More comparisons: Llama 3.3 70B vs Qwen2.5 72B · Qwen3-8B vs Qwen 2.5 7B · every Qwen2.5 72B comparison · every Qwen 2.5 7B comparison · all 662 price comparisons

Questions

Is Qwen2.5 72B cheaper than Qwen 2.5 7B?

No. On a blend of 3 input tokens to 1 output token, Qwen 2.5 7B costs USD 0.05 per 1M tokens and Qwen2.5 72B USD 0.37, at each model's cheapest standard price. For input alone, Qwen 2.5 7B is cheaper (USD 0.05 against USD 0.36 per 1M); for output alone, Qwen 2.5 7B is (USD 0.05 against USD 0.40). This compares price only, not quality.

What do Qwen2.5 72B and Qwen 2.5 7B cost per 1M tokens?

Qwen2.5 72B: from USD 0.36 input and USD 0.40 output at DeepInfra; median of 3 standard prices USD 0.37 / USD 0.40. Qwen 2.5 7B: from USD 0.05 input and USD 0.05 output at SiliconFlow; median of 3 standard prices USD 0.05 / USD 0.13. Each price is read from the provider's own pricing page; the source is linked in the table.

Which providers sell both Qwen2.5 72B and Qwen 2.5 7B?

2 providers have a standard price for both: Featherless (USD 0.3775 and USD 0.1775 at 3:1), SiliconFlow (USD 0.59 and USD 0.05 at 3:1). The first is the lowest combined price. Buying each model from its own cheapest provider costs USD 0.37 and USD 0.05.

At what input:output ratio does the cheaper of Qwen2.5 72B and Qwen 2.5 7B change?

Qwen 2.5 7B is cheaper at every mix of input and output tokens, before cache discounts: its cheapest input price and its cheapest output price are both at or below Qwen2.5 72B's, so the input:output ratio never changes the answer.

Do Qwen2.5 72B and Qwen 2.5 7B have cached input pricing?

Qwen2.5 72B: no standard price lists a cached-input discount. Qwen 2.5 7B from USD 0.136 per 1M cached input tokens (Featherless; 1 of 3 standard prices list a discount). Cached tokens are billed at the input price where no cache price is published.

Which has the larger context window, Qwen2.5 72B or Qwen 2.5 7B?

Both go up to 32,768 tokens on the providers that state it.

Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity