Qwen3.5 4B (Reasoning) API Pricing

Alibaba · released 2026-03-02 · merges 2 reasoning-effort tiers

Benchmarks shown for the highest-scoring reasoning tier. Merged tiers: Qwen3.5 4B (Reasoning), Qwen3.5 4B (Non-reasoning).

Pricing

$0.03
input / 1M tokens
< 0.01¢ / 1K
$0.15
output / 1M tokens
0.015¢ / 1K
$0.06
blended / 1M tokens
< 0.01¢ / 1K

What it actually costs

WorkloadTokensCost
Short chat reply500 in / 300 out0.006¢
RAG query4,000 in / 500 out0.0195¢
Long document summary50,000 in / 1,000 out0.165¢
Code review of a 2k-line file30,000 in / 2,000 out0.12¢
1,000 chat replies/day for 30 days30,000 × (500 in / 300 out)$1.80

Cost = (input tokens ÷ 1,000,000 × $0.03) + (output tokens ÷ 1,000,000 × $0.15).

Benchmarks

13.1
Intelligence
#202 of 533
22.6
Coding
#136 of 197
Math
18
Tokens/sec
0.59s
Time to first token

Leaderboard neighbours

ModelIntelligenceBlended $/1M
Command A+13.1
DeepSeek R1 0528 (May '25)13.1$1.763
Gemini 2.5 Flash (Reasoning)13.1$0.85
Qwen3.5 4B (Reasoning)13.1$0.06
GPT-5 nano (high)13.0$0.138
Nemotron 3.5 Lightning12.9$0.095
GLM-4.5 (Reasoning)12.8

More from Alibaba

ModelIntelligenceBlended $/1M
Qwen3.8 Max (0902)45.4$3.00
Qwen3.8 Max40.2$3.00
Qwen3.8 2.4T A95B39.9$3.00
Qwen3.8-Flash-Next39.8$0.23
Qwen3.8 27B (xhigh)33.7$1.125
Qwen3.7 Max29.5$3.75
Qwen3.6 Max Preview28.4$2.925
Qwen3.6 Plus27.0$1.125

Count tokens & estimate cost for Qwen3.5 4B (Reasoning) See Qwen3.5 4B (Reasoning) on the leaderboard

Source: Artificial Analysis · last synced 2026-09-22

FAQ

Who develops Qwen3.5 4B (Reasoning)?

Qwen3.5 4B (Reasoning) is developed by Alibaba, released 2026-03-02.

How much does Qwen3.5 4B (Reasoning) cost per million tokens?

Qwen3.5 4B (Reasoning) costs $0.03 per 1M input tokens and $0.15 per 1M output tokens ($0.06 blended), per Artificial Analysis's official pricing data.

How does Qwen3.5 4B (Reasoning) rank on benchmarks?

Qwen3.5 4B (Reasoning) scores 13.1 on the Artificial Analysis Intelligence Index, ranking #202 of 533 models we track.

Where does this data come from?

Pricing and benchmark data for Qwen3.5 4B (Reasoning) comes from Artificial Analysis (artificialanalysis.ai), synced daily. Last synced 2026-09-22.