Model Comparison
Coder-V2
vs. ERNIE 4.5 300B A47B
Comparing 2 AI models · 12 benchmarks · DeepSeek, Baidu
Recommended Pick
Ba ERNIE 4.5 300B A47B 3 metric wins Strongest on: Reasoning, Intelligence, MATH 500
Lowest Price
Coder-V2
$0.00/1M input price
Best Reasoning
ERNIE 4.5 300B A47B
45.5 reasoning score
Blends available reasoning benchmarks
Best for Coding
ERNIE 4.5 300B A47B
14.5 coding index
Composite Indices
Higher is better; speed and price are normalized
Standard Benchmarks
Only benchmarks with data are shown
Differences That Matter
Price gap
Coder-V2 is ∞x cheaper on input tokens than ERNIE 4.5 300B A47B.
Reasoning gap
ERNIE 4.5 300B A47B leads Coder-V2 by 5.8 points on reasoning.
Top-pick rationale
ERNIE 4.5 300B A47B wins 3 measurable categories, including Reasoning, Intelligence, MATH 500.
Response Face-Off
Run one prompt through the selected models and compare response quality with live speed and cost context.
Coder-V2
DeepSeek
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
ERNIE 4.5 300B A47B
Baidu
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Which answer was more useful?
Full Comparison
| Metric | De Coder-V2 | Top Pick Ba ERNIE 4.5 300B A47B |
|---|---|---|
| Pricing per 1M tokens | ||
| Input Cost | $0.00/1M | $0.28/1M |
| Output Cost | $0.00/1M | $1.10/1M |
| Blended (3:1) | — | $0.49/1M |
| Specifications | ||
| Organization | DeepSeek | Baidu |
| Release Date | Jun 17, 2024 | Jun 30, 2025 |
| Performance & Speed | ||
| Throughput | — | — |
| TTFT | — | — |
| Latency | — | — |
| Composite Indices | ||
| Value Score | — | 100.0 |
| Reasoning Score | 39.7 | 45.5 |
| Intelligence | 5.1 | 9.0 |
| Coding | — | 14.5 |
| Math | — | 41.3 |
| Standard Benchmarks | ||
| GPQA | — | 81.1% |
| MMLU Pro | — | 77.6% |
| HLE | — | 3.5% |
| LiveCodeBench | — | 46.7% |
| MATH 500 | 74.3% | 93.1% |
| AIME 2025 | — | 41.3% |
| AIME (Original) | — | 49.3% |
| SciCode | — | 31.5% |
| LCR | — | 2.3% |
| IFBench | — | 39.1% |
| TAU-bench v2 | — | 0.0% |
| TerminalBench Hard | — | 6.1% |
Key Takeaways
Coder-V2 offers the best value at $0.00/1M, making it ideal for high-volume applications and cost-conscious projects.
ERNIE 4.5 300B A47B has the strongest reasoning profile with a 45.5 reasoning score, combining the available reasoning-heavy benchmarks.
ERNIE 4.5 300B A47B reaches a 14.5 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
Coder-V2
- Cost-sensitive applications
- High-volume processing
ERNIE 4.5 300B A47B
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
