Key Takeaways
GPT-OSS-120b wins:
- Cheaper input tokens
- Faster response time
- Higher intelligence benchmark
- Better at coding
- Better at math
Qwen3 235B A22B Instruct 2507 wins:
- Cheaper output tokens
- Larger context window
- Supports tool calls
Price Advantage
GPT-OSS-120b
Benchmark Advantage
GPT-OSS-120b
Context Window
Qwen3 235B A22B Instruct 2507
Speed
GPT-OSS-120b
Pricing Comparison
Price Comparison
| Metric | GPT-OSS-120b | Qwen3 235B A22B Instruct 2507 | Winner |
|---|---|---|---|
| Input (per 1M tokens) | $0.04 | $0.07 | GPT-OSS-120b |
| Output (per 1M tokens) | $0.19 | $0.10 | Qwen3 235B A22B Instruct 2507 |
Using a 3:1 input/output ratio, GPT-OSS-120b is 2% cheaper overall.
GPT-OSS-120b Providers
Chutes $0.04 (Cheapest)
SiliconFlow $0.05
Novita $0.05
Clarifai $0.09
Google $0.09
Qwen3 235B A22B Instruct 2507 Providers
DeepInfra $0.07 (Cheapest)
Chutes $0.08
SiliconFlow $0.09
Novita $0.09
Parasail $0.10
Benchmark Comparison
8
Benchmarks Compared
5
GPT-OSS-120b Wins
1
Qwen3 235B A22B Instruct 2507 Wins
Benchmark Scores
| Benchmark | GPT-OSS-120b | Qwen3 235B A22B Instruct 2507 | Winner |
|---|---|---|---|
Intelligence Index Overall intelligence score | 33.3 | 24.7 | |
Coding Index Code generation & understanding | 28.6 | 22.1 | |
Math Index Mathematical reasoning | 93.4 | 71.7 | |
MMLU Pro Academic knowledge | 80.8 | 82.8 | |
GPQA Graduate-level science | 78.2 | 75.3 | |
LiveCodeBench Competitive programming | 87.8 | 52.4 | |
Aider Real-world code editing | 41.8 | - | - |
AIME Competition math | - | 71.7 | - |
GPT-OSS-120b significantly outperforms in coding benchmarks.
Cost vs Quality
X-axis:
Y-axis:
Loading chart...
GPT-OSS-120b
Other models
Context & Performance
Context Window
GPT-OSS-120b
131,072
tokens
Qwen3 235B A22B Instruct 2507
262,144
tokens
Qwen3 235B A22B Instruct 2507 has a 50% larger context window.
Speed Performance
| Metric | GPT-OSS-120b | Qwen3 235B A22B Instruct 2507 | Winner |
|---|---|---|---|
| Tokens/second | 311.5 tok/s | 53.5 tok/s | |
| Time to First Token | 0.47s | 1.15s |
GPT-OSS-120b responds 482% faster on average.
Capabilities
Feature Comparison
| Feature | GPT-OSS-120b | Qwen3 235B A22B Instruct 2507 |
|---|---|---|
| Vision (Image Input) | ||
| Tool/Function Calls | ||
| Reasoning Mode | ||
| Audio Input | ||
| Audio Output | ||
| PDF Input | ||
| Prompt Caching | ||
| Web Search |
License & Release
| Property | GPT-OSS-120b | Qwen3 235B A22B Instruct 2507 |
|---|---|---|
| License | Open Source | Open Source |
| Author | OpenAI | Qwen |
| Released | Aug 2025 | Jul 2025 |
GPT-OSS-120b Modalities
Input
text
Output
text
Qwen3 235B A22B Instruct 2507 Modalities
Input
text
Output
text