Models, measured side by side
Phi-4 vs Qwen3 Coder 30B A3B Instruct
Compare Phi-4, Qwen3 Coder 30B A3B Instruct by token pricing, context length, benchmark results, speed and tool support.
Monthly workload
Monthly cost breakdown
InputOutput
Phi-4
$2.10 / mo
Qwen3 Coder 30B A3B Instruct
$2.80 / mo
Performance comparison
Phi-4
5.9
Qwen3 Coder 30B A3B Instruct
9.6
Overview
Pricing & capacity
| Pricing & capacity | Phi-4 | Qwen3 Coder 30B A3B Instruct |
|---|---|---|
| Input / 1M tokens | $0.07 | $0.07 |
| Output / 1M tokens | $0.14 | $0.28 |
| Monthly cost | $2.10 | $2.80 |
| Context window | 16K | 262K |
| Maximum output | 15K | 236K |
Input / 1M tokens
- Phi-4
- $0.07
- Qwen3 Coder 30B A3B Instruct
- $0.07
Output / 1M tokens
- Phi-4
- $0.14
- Qwen3 Coder 30B A3B Instruct
- $0.28
Monthly cost
- Phi-4
- $2.10
- Qwen3 Coder 30B A3B Instruct
- $2.80
Context window
- Phi-4
- 16K
- Qwen3 Coder 30B A3B Instruct
- 262K
Maximum output
- Phi-4
- 15K
- Qwen3 Coder 30B A3B Instruct
- 236K
Benchmarks & performance
| Benchmarks & performance | Phi-4 | Qwen3 Coder 30B A3B Instruct |
|---|---|---|
| Arena Elo | 1,123 | Not listed |
| MMLU-Pro | 60.7% | Not listed |
| Intelligence Index | 5.9 | 9.6 |
| GPQA | 57.5% | 51.6% |
| Humanity’s Last Exam | 3.8% | 3.8% |
| Output speed | 39 tokens/s | 87 tokens/s |
| Time to first token | 2.59 s | 2.65 s |
Arena Elo
- Phi-4
- 1,123
- Qwen3 Coder 30B A3B Instruct
- Not listed
MMLU-Pro
- Phi-4
- 60.7%
- Qwen3 Coder 30B A3B Instruct
- Not listed
Intelligence Index
- Phi-4
- 5.9
- Qwen3 Coder 30B A3B Instruct
- 9.6
GPQA
- Phi-4
- 57.5%
- Qwen3 Coder 30B A3B Instruct
- 51.6%
Humanity’s Last Exam
- Phi-4
- 3.8%
- Qwen3 Coder 30B A3B Instruct
- 3.8%
Output speed
- Phi-4
- 39 tokens/s
- Qwen3 Coder 30B A3B Instruct
- 87 tokens/s
Time to first token
- Phi-4
- 2.59 s
- Qwen3 Coder 30B A3B Instruct
- 2.65 s
Tools & features
| Tools & features | Phi-4 | Qwen3 Coder 30B A3B Instruct |
|---|---|---|
| Function calling | Not listed | Supported |
| Structured outputs | Supported | Supported |
| JSON mode | Supported | Supported |
| Log probabilities | Not listed | Supported |
| Deterministic seed | Supported | Supported |
Function calling
- Phi-4
- Not listed
- Qwen3 Coder 30B A3B Instruct
- Supported
Structured outputs
- Phi-4
- Supported
- Qwen3 Coder 30B A3B Instruct
- Supported
JSON mode
- Phi-4
- Supported
- Qwen3 Coder 30B A3B Instruct
- Supported
Log probabilities
- Phi-4
- Not listed
- Qwen3 Coder 30B A3B Instruct
- Supported
Deterministic seed
- Phi-4
- Supported
- Qwen3 Coder 30B A3B Instruct
- Supported
API & availability
| API & availability | Phi-4 | Qwen3 Coder 30B A3B Instruct |
|---|---|---|
| API identifier | phi-4 | qwen3-coder-30b-a3b-instruct |
| API providers | DeepInfra | Not listed |
| Knowledge cutoff | 2024-06-30 | 2025-06-30 |
| Prices checked | Oct 8, 2026 | Oct 8, 2026 |
API identifier
- Phi-4
phi-4- Qwen3 Coder 30B A3B Instruct
qwen3-coder-30b-a3b-instruct
API providers
- Phi-4
- DeepInfra
- Qwen3 Coder 30B A3B Instruct
- Not listed
Knowledge cutoff
- Phi-4
- 2024-06-30
- Qwen3 Coder 30B A3B Instruct
- 2025-06-30
Prices checked
- Phi-4
- Oct 8, 2026
- Qwen3 Coder 30B A3B Instruct
- Oct 8, 2026
About the models
Qwen3 Coder 30B A3B Instruct
Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use. Built on the...
Full pricing & detailsComparison FAQ
Phi-4: $0.07 input and $0.14 output per million tokens. Qwen3 Coder 30B A3B Instruct: $0.07 input and $0.28 output per million tokens. The cheaper choice depends on your input-to-output ratio; a model can have a lower input rate but a higher output rate.