Model Comparison
LFM2.5-1.2B-Instruct
vs. MiniCPM-V 4.6 1.3B
Comparing 2 AI models · 7 benchmarks · Liquid AI, OpenBMB
Recommended Pick
Li LFM2.5-1.2B-Instruct 5 metric wins Strongest on: Coding, GPQA, HLE
Lowest Price
LFM2.5-1.2B-Instruct
$0.00/1M input price
Best Reasoning
MiniCPM-V 4.6 1.3B
16.0 reasoning score
Blends available reasoning benchmarks
Best for Coding
LFM2.5-1.2B-Instruct
0.8 coding index
Composite Indices
Higher is better; speed and price are normalized
Standard Benchmarks
Only benchmarks with data are shown
Differences That Matter
Reasoning gap
MiniCPM-V 4.6 1.3B leads LFM2.5-1.2B-Instruct by 0.2 points on reasoning.
Coding gap
LFM2.5-1.2B-Instruct leads MiniCPM-V 4.6 1.3B by 0.1 points on coding.
Top-pick rationale
LFM2.5-1.2B-Instruct wins 5 measurable categories, including Coding, GPQA, HLE, SciCode.
Response Face-Off
Run one prompt through the selected models and compare response quality with live speed and cost context.
LFM2.5-1.2B-Instruct
Liquid AI
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
MiniCPM-V 4.6 1.3B
OpenBMB
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Which answer was more useful?
Full Comparison
| Metric | Top Pick Li LFM2.5-1.2B-Instruct | Op MiniCPM-V 4.6 1.3B |
|---|---|---|
| Pricing per 1M tokens | ||
| Input Cost | $0.00/1M | $0.00/1M |
| Output Cost | $0.00/1M | $0.00/1M |
| Specifications | ||
| Organization | Liquid AI | OpenBMB |
| Release Date | Jan 5, 2026 | May 11, 2026 |
| Performance & Speed | ||
| Throughput | — | — |
| TTFT | — | — |
| Latency | — | — |
| Composite Indices | ||
| Reasoning Score | 15.8 | 16.0 |
| Intelligence | 8.0 | 12.7 |
| Coding | 0.8 | 0.7 |
| Standard Benchmarks | ||
| GPQA | 32.6% | 30.5% |
| HLE | 6.8% | 4.9% |
| SciCode | 2.3% | 2.1% |
| LCR | 0.0% | 6.3% |
| IFBench | 43.8% | 26.7% |
| TAU-bench v2 | 10.8% | 87.7% |
| TerminalBench Hard | 0.0% | 0.0% |
Key Takeaways
LFM2.5-1.2B-Instruct offers the best value at $0.00/1M, making it ideal for high-volume applications and cost-conscious projects.
MiniCPM-V 4.6 1.3B has the strongest reasoning profile with a 16.0 reasoning score, combining the available reasoning-heavy benchmarks.
LFM2.5-1.2B-Instruct reaches a 0.8 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
LFM2.5-1.2B-Instruct
- Cost-sensitive applications
- High-volume processing
- Code generation
- Software development
MiniCPM-V 4.6 1.3B
- Complex reasoning tasks
- Research & analysis
