The AI arena is free today

Open Superagent

Hy4 flash vs MiniCPM-SALA

Hy4 flash leads the LLM Stats Score 40.8 to 13.9.

Tencent · OpenBMB · Updated for 2026

Which is better?

Hy4 flash leads the overall LLM Stats Score 40.8 to 13.9, ranking #84 overall.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose Hy4 flash

  • overall performance matters — it scores 40.8 and ranks #84 on LLM Stats
  • your work emphasizes reasoning — it leads those capability indexes
  • you want the most recent training data — it shipped Nov 2026

Choose MiniCPM-SALA

  • you need open weights you can self-host or fine-tune

At a glance

The differences that matter most.

Core performance indexes
40.8
#84
13.9
#285
40.4
#77
15.2
#268
23.5
#107
20.8
#124
Cost, coverage & limits
Benchmark wins
—
—
Input price
$0.07 / M
— / M
Output price
$0.28 / M
— / M
Context window
960,000
—

Individual benchmarks

9 reported for Hy4 flash · 24 for MiniCPM-SALA

No common benchmarks found

Hy4 flash and MiniCPM-SALAdon't have any common benchmark datasets to compare. They may have been evaluated on different testing suites.

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

214.5B diff

Hy4 flash has 214.5B more parameters than MiniCPM-SALA, making it 2263.6% larger.

Tencent
Hy4 flash
224.0Bparameters
OpenBMB
MiniCPM-SALA
9.5Bparameters
224.0B
Hy4 flash
9.5B
MiniCPM-SALA

Context Window

Maximum input and output token capacity

Only Hy4 flash specifies input context (960,000 tokens). Only Hy4 flash specifies output context (64,000 tokens).

Tencent
Hy4 flash
Input960,000 tokens
Output64,000 tokens
OpenBMB
MiniCPM-SALA
Input- tokens
Output- tokens
Mon Nov 16 2026 • llm-stats.com

Input capabilities

Documented input modalities across available providers

Hy4 flash supports multimodal inputs, whereas MiniCPM-SALA does not.

Hy4 flash can handle both text and other forms of data like images, making it suitable for multimodal applications.

Hy4 flash

Text
Images
Audio
Video

MiniCPM-SALA

Text
Images
Audio
Video

Release Timeline

When each model was launched

Hy4 flash was released on 2026-11-16, while MiniCPM-SALA was released on 2026-02-11.

Hy4 flash is 9 months newer than MiniCPM-SALA.

Hy4 flash

Nov 16, 2026

0 days ago

9mo newer
MiniCPM-SALA

Feb 11, 2026

9 months ago

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?

Judge for yourself.

Run your own prompts against Hy4 flash and MiniCPM-SALA side-by-side, then vote on the output you prefer.

Hy4 flash
✓ Preferred
MiniCPM-SALA
Open in Playground

FAQ

Common questions about Hy4 flash vs MiniCPM-SALA.

Which is better, Hy4 flash or MiniCPM-SALA?

Hy4 flash leads the LLM Stats Score 40.8 to 13.9. Hy4 flash is made by Tencent and MiniCPM-SALA is made by OpenBMB. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does Hy4 flash compare to MiniCPM-SALA in benchmarks?

Hy4 flash scores MMMU-Pro: 75.8%, OSWorld 2.1 (partial): 68.1%, DeepSWE 1.1: 64.6%, APEX-Agents: 63.1%, Humanity's Last Exam (with tools): 55.2%. MiniCPM-SALA scores HumanEval: 95.1%, RULER 64k: 92.7%, RULER 128k: 89.4%, MBPP: 89.1%, RULER 512K: 87.1%.

What are the context window sizes for Hy4 flash and MiniCPM-SALA?

Hy4 flash supports 960K tokens and MiniCPM-SALA supports an unknown number of tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Hy4 flash and MiniCPM-SALA?

Key differences include LLM Stats Score (40.8 vs 13.9), multimodal support (yes vs no), licensing (Unknown vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Who makes Hy4 flash and MiniCPM-SALA?

Hy4 flash is developed by Tencent and MiniCPM-SALA is developed by OpenBMB.