# DeepSeek V3 (671B MoE) — AI Tool Profile
> Source: https://agentrobbies.com/tools/deepseek-v3
> Ecosystem: DEEPSEEK | Verified: 2026-10-11
> 100% Independent Listing • Zero Affiliate Links

## Description
High-throughput Mixture-of-Experts (MoE) foundation model activating 37B parameters per token with Multi-head Latent Attention (MLA).

## Access & Pricing
Free on web/app; API at $0.14/1M input tokens.

## Key Capabilities
- 60 tokens/second inference throughput on modern clusters
- 128k context window with perfect retrieval accuracy
- Leading open-weight benchmark on MMLU and coding datasets

## Alternatives & Replaces
- Llama 3.3 70B
- GPT-4o

## Robbie's Take
**Robbie's Performance Take:** Ideal for latency-critical agent loops where fast token generation directly reduces user wait times. Benchmark output quality on your specific task before swapping out larger frontier reasoning models.

## Direct Link
https://github.com/deepseek-ai/DeepSeek-V3