Model Comparison
Ministral 3 8B
vs. Large 3
Comparing 2 AI models · 10 benchmarks · Mistral
Recommended Pick
Mi Large 3 12 metric wins Strongest on: Reasoning, Intelligence, Coding
Best Value
Ministral 3 8B
100.0 value score
24.7 reasoning / $0.15/1M
Lowest Price
Ministral 3 8B
$0.15/1M input price
Best Reasoning
Large 3
32.9 reasoning score
Blends available reasoning benchmarks
Best for Coding
Large 3
22.7 coding index
Composite Indices
Higher is better; speed and price are normalized
Standard Benchmarks
Only benchmarks with data are shown
Differences That Matter
Best value
Ministral 3 8B has the strongest quality-to-price mix at 100.0 out of 100 value points.
Price gap
Ministral 3 8B is 3.3x cheaper on input tokens than Large 3.
Speed gap
Ministral 3 8B generates about 1.6x as many tokens per second as Large 3.
Reasoning gap
Large 3 leads Ministral 3 8B by 8.1 points on reasoning.
Coding gap
Large 3 leads Ministral 3 8B by 12.7 points on coding.
Response Face-Off
Run one prompt through the selected models and compare response quality with live speed and cost context.
Ministral 3 8B
Mistral
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Large 3
Mistral
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Which answer was more useful?
Full Comparison
| Metric | Mi Ministral 3 8B | Top Pick Mi Large 3 |
|---|---|---|
| Pricing per 1M tokens | ||
| Input Cost | $0.15/1M | $0.50/1M |
| Output Cost | $0.15/1M | $1.50/1M |
| Blended (3:1) | $0.15/1M | $0.75/1M |
| Specifications | ||
| Organization | Mistral | Mistral |
| Release Date | Dec 2, 2025 | Dec 2, 2025 |
| Performance & Speed | ||
| Throughput | 103.7 tok/s | 63.0 tok/s |
| TTFT | 446ms | 674ms |
| Latency | 446ms | 674ms |
| Composite Indices | ||
| Value Score | 100.0 | 26.6 |
| Reasoning Score | 24.7 | 32.9 |
| Intelligence | 8.9 | 16.2 |
| Coding | 10.0 | 22.7 |
| Math | 31.7 | 38.0 |
| Standard Benchmarks | ||
| GPQA | 47.1% | 68.0% |
| MMLU Pro | 64.2% | 80.7% |
| HLE | 4.3% | 4.1% |
| LiveCodeBench | 30.3% | 46.5% |
| AIME 2025 | 31.7% | 38.0% |
| SciCode | 20.8% | 36.2% |
| LCR | 24.0% | 34.7% |
| IFBench | 29.1% | 36.2% |
| TAU-bench v2 | 26.6% | 24.6% |
| TerminalBench Hard | 4.5% | 15.9% |
Key Takeaways
Ministral 3 8B offers the best value at $0.15/1M, making it ideal for high-volume applications and cost-conscious projects.
Large 3 has the strongest reasoning profile with a 32.9 reasoning score, combining the available reasoning-heavy benchmarks.
Large 3 reaches a 22.7 coding index, making it the top choice for software development and code generation tasks.
All models support context windows of ∞+ tokens, suitable for processing lengthy documents and maintaining extended conversations.
When to Choose Each Model
Ministral 3 8B
- Cost-sensitive applications
- High-volume processing
Large 3
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
