Model Comparison

DeepSeek-V2.5 vs Llama 3.2 90B Instruct

Both models are evenly matched across the benchmarks. DeepSeek-V2.5 is 2.1x cheaper per token.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

2 benchmarks

DeepSeek-V2.5 outperforms in 1 benchmarks (MATH), while Llama 3.2 90B Instruct is better at 1 benchmark (MMLU).

Both models are evenly matched across the benchmarks.

Wed Apr 01 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

DeepSeek-V2.5 costs less

For input processing, DeepSeek-V2.5 ($0.14/1M tokens) is 2.5x cheaper than Llama 3.2 90B Instruct ($0.35/1M tokens).

For output processing, DeepSeek-V2.5 ($0.28/1M tokens) is 1.4x cheaper than Llama 3.2 90B Instruct ($0.40/1M tokens).

In conclusion, Llama 3.2 90B Instruct is more expensive than DeepSeek-V2.5.*

* Using a 3:1 ratio of input to output tokens

Lowest available price from all providers
Wed Apr 01 2026 • llm-stats.com
DeepSeek
DeepSeek-V2.5
Input tokens$0.14
Output tokens$0.28
Best providerDeepSeek
Meta
Llama 3.2 90B Instruct
Input tokens$0.35
Output tokens$0.40
Best providerDeepinfra
Notice missing or incorrect data?Start an Issue→

Model Size

Parameter count comparison

146.0B diff

DeepSeek-V2.5 has 146.0B more parameters than Llama 3.2 90B Instruct, making it 162.2% larger.

DeepSeek
DeepSeek-V2.5
236.0Bparameters
Meta
Llama 3.2 90B Instruct
90.0Bparameters
236.0B
DeepSeek-V2.5
90.0B
Llama 3.2 90B Instruct

Context Window

Maximum input and output token capacity

Llama 3.2 90B Instruct accepts 128,000 input tokens compared to DeepSeek-V2.5's 8,192 tokens. Llama 3.2 90B Instruct can generate longer responses up to 128,000 tokens, while DeepSeek-V2.5 is limited to 8,192 tokens.

DeepSeek
DeepSeek-V2.5
Input8,192 tokens
Output8,192 tokens
Meta
Llama 3.2 90B Instruct
Input128,000 tokens
Output128,000 tokens
Wed Apr 01 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Llama 3.2 90B Instruct supports multimodal inputs, whereas DeepSeek-V2.5 does not.

Llama 3.2 90B Instruct can handle both text and other forms of data like images, making it suitable for multimodal applications.

DeepSeek-V2.5

Text
Images
Audio
Video

Llama 3.2 90B Instruct

Text
Images
Audio
Video

License

Usage and distribution terms

DeepSeek-V2.5 is licensed under deepseek, while Llama 3.2 90B Instruct uses Llama 3.2.

License differences may affect how you can use these models in commercial or open-source projects.

DeepSeek-V2.5

deepseek

Open weights

Llama 3.2 90B Instruct

Llama 3.2

Open weights

Release Timeline

When each model was launched

DeepSeek-V2.5 was released on 2024-05-08, while Llama 3.2 90B Instruct was released on 2024-09-25.

Llama 3.2 90B Instruct is 5 months newer than DeepSeek-V2.5.

DeepSeek-V2.5

May 8, 2024

1.9 years ago

Llama 3.2 90B Instruct

Sep 25, 2024

1.5 years ago

4mo newer

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

Provider Availability

DeepSeek-V2.5 is available from DeepSeek, DeepInfra, Hyperbolic. Llama 3.2 90B Instruct is available from DeepInfra, Bedrock, Fireworks, Together, Hyperbolic.

DeepSeek-V2.5

deepseek logo
DeepSeek
Input Price:Input: $0.14/1MOutput Price:Output: $0.28/1M
deepinfra logo
Deepinfra
Input Price:Input: $0.70/1MOutput Price:Output: $1.40/1M
hyperbolic logo
Hyperbolic
Input Price:Input: $2.00/1MOutput Price:Output: $2.00/1M

Llama 3.2 90B Instruct

deepinfra logo
Deepinfra
Input Price:Input: $0.35/1MOutput Price:Output: $0.40/1M
bedrock logo
AWS Bedrock
Input Price:Input: $0.72/1MOutput Price:Output: $0.72/1M
fireworks logo
Fireworks
Input Price:Input: $0.89/1MOutput Price:Output: $0.89/1M
together logo
Together
Input Price:Input: $1.20/1MOutput Price:Output: $1.20/1M
hyperbolic logo
Hyperbolic
Input Price:Input: $2.00/1MOutput Price:Output: $2.00/1M
* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

DeepSeek
Less expensive input tokens
Less expensive output tokens
Higher MATH score (74.7% vs 68.0%)
Meta
Larger context window (128,000 tokens)
Supports multimodal inputs
Higher MMLU score (86.0% vs 80.4%)

Detailed Comparison

AI Model Comparison Table
Feature
DeepSeek
DeepSeek-V2.5
Meta
Llama 3.2 90B Instruct

FAQ

Common questions about DeepSeek-V2.5 vs Llama 3.2 90B Instruct

Both models are evenly matched across the benchmarks. DeepSeek-V2.5 is made by DeepSeek and Llama 3.2 90B Instruct is made by Meta. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.
DeepSeek-V2.5 scores GSM8k: 95.1%, MT-Bench: 90.2%, HumanEval: 89.0%, BBH: 84.3%, AlignBench: 80.4%. Llama 3.2 90B Instruct scores AI2D: 92.3%, DocVQA: 90.1%, MGSM: 86.9%, MMLU: 86.0%, ChartQA: 85.5%.
DeepSeek-V2.5 is 2.5x cheaper for input tokens. DeepSeek-V2.5 costs $0.14/M input and $0.28/M output via deepseek. Llama 3.2 90B Instruct costs $0.35/M input and $0.40/M output via deepinfra.
DeepSeek-V2.5 supports 8K tokens and Llama 3.2 90B Instruct supports 128K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.
Key differences include context window (8K vs 128K), input pricing ($0.14 vs $0.35/M), multimodal support (no vs yes), licensing (deepseek vs Llama 3.2). See the full comparison above for benchmark-by-benchmark results.
DeepSeek-V2.5 is developed by DeepSeek and Llama 3.2 90B Instruct is developed by Meta.