Key Takeaways
Devstral 2 2512 wins:
- Larger context window
- Supports tool calls
GPT-OSS-120b wins:
- Cheaper input tokens
- Cheaper output tokens
- Faster response time
- Higher intelligence benchmark
- Better at coding
- Better at math
Price Advantage
GPT-OSS-120b
Benchmark Advantage
GPT-OSS-120b
Context Window
Devstral 2 2512
Speed
GPT-OSS-120b
Pricing Comparison
Price Comparison
| Metric | Devstral 2 2512 | GPT-OSS-120b | Winner |
|---|---|---|---|
| Input (per 1M tokens) | $0.05 | $0.04 | GPT-OSS-120b |
| Output (per 1M tokens) | $0.22 | $0.19 | GPT-OSS-120b |
Using a 3:1 input/output ratio, GPT-OSS-120b is 17% cheaper overall.
Devstral 2 2512 Providers
Chutes $0.05 (Cheapest)
Mistral $0.40
GPT-OSS-120b Providers
DeepInfra $0.04 (Cheapest)
Chutes $0.04
SiliconFlow $0.05
NCompass $0.05
Novita $0.05
Benchmark Comparison
7
Benchmarks Compared
0
Devstral 2 2512 Wins
6
GPT-OSS-120b Wins
Benchmark Scores
| Benchmark | Devstral 2 2512 | GPT-OSS-120b | Winner |
|---|---|---|---|
Intelligence Index Overall intelligence score | 22.0 | 33.3 | |
Coding Index Code generation & understanding | 23.7 | 28.6 | |
Math Index Mathematical reasoning | 36.7 | 93.4 | |
MMLU Pro Academic knowledge | 76.2 | 80.8 | |
GPQA Graduate-level science | 59.4 | 78.2 | |
LiveCodeBench Competitive programming | 44.8 | 87.8 | |
Aider Real-world code editing | - | 41.8 | - |
GPT-OSS-120b shows stronger mathematical reasoning abilities.
Cost vs Quality
X-axis:
Y-axis:
Loading chart...
Devstral 2 2512
Other models
Context & Performance
Context Window
Devstral 2 2512
262,144
tokens
Max output: 65,536 tokens
GPT-OSS-120b
131,072
tokens
Devstral 2 2512 has a 50% larger context window.
Speed Performance
| Metric | Devstral 2 2512 | GPT-OSS-120b | Winner |
|---|---|---|---|
| Tokens/second | 64.0 tok/s | 321.3 tok/s | |
| Time to First Token | 0.41s | 0.40s |
GPT-OSS-120b responds 402% faster on average.
Capabilities
Feature Comparison
| Feature | Devstral 2 2512 | GPT-OSS-120b |
|---|---|---|
| Vision (Image Input) | ||
| Tool/Function Calls | ||
| Reasoning Mode | ||
| Audio Input | ||
| Audio Output | ||
| PDF Input | ||
| Prompt Caching | ||
| Web Search |
License & Release
| Property | Devstral 2 2512 | GPT-OSS-120b |
|---|---|---|
| License | Open Source | Open Source |
| Author | Mistral AI | OpenAI |
| Released | Dec 2025 | Aug 2025 |
Devstral 2 2512 Modalities
Input
text
Output
text
GPT-OSS-120b Modalities
Input
text
Output
text