Model Comparison

DeepSeek-V3.1 vs Llama 3.2 90B Instruct

DeepSeek-V3.1 significantly outperforms across most benchmarks. Llama 3.2 90B Instruct is 1.2x cheaper per token.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

1 benchmarks

DeepSeek-V3.1 outperforms in 1 benchmarks (GPQA), while Llama 3.2 90B Instruct is better at 0 benchmarks.

DeepSeek-V3.1 significantly outperforms across most benchmarks.

Wed Apr 22 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

Llama 3.2 90B Instruct costs less

For input processing, DeepSeek-V3.1 ($0.27/1M tokens) is 1.3x cheaper than Llama 3.2 90B Instruct ($0.35/1M tokens).

For output processing, DeepSeek-V3.1 ($1.00/1M tokens) is 2.5x more expensive than Llama 3.2 90B Instruct ($0.40/1M tokens).

In conclusion, DeepSeek-V3.1 is more expensive than Llama 3.2 90B Instruct.*

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

Lowest available price from all providers
Wed Apr 22 2026 • llm-stats.com
DeepSeek
DeepSeek-V3.1
Input tokens$0.27
Output tokens$1.00
Best providerDeepinfra
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

581.0B diff

DeepSeek-V3.1 has 581.0B more parameters than Llama 3.2 90B Instruct, making it 645.6% larger.

DeepSeek
DeepSeek-V3.1
671.0Bparameters
Meta
Llama 3.2 90B Instruct
90.0Bparameters
671.0B
DeepSeek-V3.1
90.0B
Llama 3.2 90B Instruct

Context Window

Maximum input and output token capacity

DeepSeek-V3.1 accepts 163,840 input tokens compared to Llama 3.2 90B Instruct's 128,000 tokens. DeepSeek-V3.1 can generate longer responses up to 163,840 tokens, while Llama 3.2 90B Instruct is limited to 128,000 tokens.

DeepSeek
DeepSeek-V3.1
Input163,840 tokens
Output163,840 tokens
Meta
Llama 3.2 90B Instruct
Input128,000 tokens
Output128,000 tokens
Wed Apr 22 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Llama 3.2 90B Instruct supports multimodal inputs, whereas DeepSeek-V3.1 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-V3.1

Text
Images
Audio
Video

Llama 3.2 90B Instruct

Text
Images
Audio
Video

License

Usage and distribution terms

DeepSeek-V3.1 is licensed under MIT, 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-V3.1

MIT

Open weights

Llama 3.2 90B Instruct

Llama 3.2

Open weights

Release Timeline

When each model was launched

DeepSeek-V3.1 was released on 2025-01-10, while Llama 3.2 90B Instruct was released on 2024-09-25.

DeepSeek-V3.1 is 4 months newer than Llama 3.2 90B Instruct.

DeepSeek-V3.1

Jan 10, 2025

1.3 years ago

3mo newer
Llama 3.2 90B Instruct

Sep 25, 2024

1.6 years 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

Provider Availability

DeepSeek-V3.1 is available from DeepInfra, Novita. Llama 3.2 90B Instruct is available from DeepInfra, Bedrock, Fireworks, Together, Hyperbolic.

DeepSeek-V3.1

deepinfra logo
Deepinfra
Input Price:Input: $0.27/1MOutput Price:Output: $1.00/1M
novita logo
Novita
Input Price:Input: $0.27/1MOutput Price:Output: $1.00/1M

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
Larger context window (163,840 tokens)
Less expensive input tokens
Higher GPQA score (74.9% vs 46.7%)
Meta
Supports multimodal inputs
Less expensive output tokens

Detailed Comparison

AI Model Comparison Table
Feature
DeepSeek
DeepSeek-V3.1
Meta
Llama 3.2 90B Instruct

FAQ

Common questions about DeepSeek-V3.1 vs Llama 3.2 90B Instruct

DeepSeek-V3.1 significantly outperforms across most benchmarks. DeepSeek-V3.1 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-V3.1 scores SimpleQA: 93.4%, MMLU-Redux: 91.8%, MMLU-Pro: 83.7%, GPQA: 74.9%, CodeForces: 69.7%. Llama 3.2 90B Instruct scores AI2D: 92.3%, DocVQA: 90.1%, MGSM: 86.9%, MMLU: 86.0%, ChartQA: 85.5%.
DeepSeek-V3.1 is 1.3x cheaper for input tokens. DeepSeek-V3.1 costs $0.27/M input and $1.00/M output via deepinfra. Llama 3.2 90B Instruct costs $0.35/M input and $0.40/M output via deepinfra.
DeepSeek-V3.1 supports 164K 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 (164K vs 128K), input pricing ($0.27 vs $0.35/M), multimodal support (no vs yes), licensing (MIT vs Llama 3.2). See the full comparison above for benchmark-by-benchmark results.
DeepSeek-V3.1 is developed by DeepSeek and Llama 3.2 90B Instruct is developed by Meta.