Key Takeaways
MiniMax M2.5 wins:
- Faster response time
- Higher intelligence benchmark
- Better at coding
Qwen3 Coder 480B A35B (exacto) wins:
- Cheaper input tokens
- Cheaper output tokens
- Larger context window
- Better at math
- Supports tool calls
Price Advantage
Qwen3 Coder 480B A35B (exacto)
Benchmark Advantage
MiniMax M2.5
Context Window
Qwen3 Coder 480B A35B (exacto)
Speed
MiniMax M2.5
Pricing Comparison
Price Comparison
| Metric | MiniMax M2.5 | Qwen3 Coder 480B A35B (exacto) | Winner |
|---|---|---|---|
| Input (per 1M tokens) | $0.30 | $0.22 | Qwen3 Coder 480B A35B (exacto) |
| Output (per 1M tokens) | $1.20 | $1.00 | Qwen3 Coder 480B A35B (exacto) |
| Cache Read (per 1M) | $30000.00 | $22000.00 | Qwen3 Coder 480B A35B (exacto) |
Using a 3:1 input/output ratio, Qwen3 Coder 480B A35B (exacto) is 21% cheaper overall.
MiniMax M2.5 Providers
Minimax $0.30 (Cheapest)
Novita $0.30 (Cheapest)
Qwen3 Coder 480B A35B (exacto) Providers
Google $0.22 (Cheapest)
SiliconFlow $0.25
Novita $0.30
Vercel $0.40
Nebius $0.40
Benchmark Comparison
7
Benchmarks Compared
3
MiniMax M2.5 Wins
0
Qwen3 Coder 480B A35B (exacto) Wins
Benchmark Scores
| Benchmark | MiniMax M2.5 | Qwen3 Coder 480B A35B (exacto) | Winner |
|---|---|---|---|
Intelligence Index Overall intelligence score | 42.0 | 24.6 | |
Coding Index Code generation & understanding | 37.4 | 24.6 | |
Math Index Mathematical reasoning | - | 39.3 | - |
MMLU Pro Academic knowledge | - | 78.8 | - |
GPQA Graduate-level science | 84.8 | 61.8 | |
LiveCodeBench Competitive programming | - | 58.5 | - |
AIME Competition math | - | 47.7 | - |
MiniMax M2.5 significantly outperforms in coding benchmarks.
Cost vs Quality
X-axis:
Y-axis:
Loading chart...
Other models
Context & Performance
Context Window
MiniMax M2.5
204,800
tokens
Max output: 131,072 tokens
Qwen3 Coder 480B A35B (exacto)
262,144
tokens
Qwen3 Coder 480B A35B (exacto) has a 22% larger context window.
Speed Performance
| Metric | MiniMax M2.5 | Qwen3 Coder 480B A35B (exacto) | Winner |
|---|---|---|---|
| Tokens/second | 71.0 tok/s | 54.4 tok/s | |
| Time to First Token | 1.67s | 1.67s | - |
MiniMax M2.5 responds 30% faster on average.
Capabilities
Feature Comparison
| Feature | MiniMax M2.5 | Qwen3 Coder 480B A35B (exacto) |
|---|---|---|
| Vision (Image Input) | ||
| Tool/Function Calls | ||
| Reasoning Mode | ||
| Audio Input | ||
| Audio Output | ||
| PDF Input | ||
| Prompt Caching | ||
| Web Search |
License & Release
| Property | MiniMax M2.5 | Qwen3 Coder 480B A35B (exacto) |
|---|---|---|
| License | Proprietary | Open Source |
| Author | Minimax | Qwen |
| Released | Feb 2026 | Jul 2025 |
MiniMax M2.5 Modalities
Input
text
Output
text
Qwen3 Coder 480B A35B (exacto) Modalities
Input
text
Output
text