DeepSeek
DeepSeek V4 Pro 0813
Open weights
- Intelligence
- 20.4
- Rank 242 of 744 measured models
- Output speed
- 91.2 t/s
- Rank 194 of 403 measured models
- Price/1M
- $1.98
- Rank 463 of 659 measured models
- Context window
- 1M
- Rank 84 of 1199 measured models
- Cost per task
- $0.356
- Rank 92 of 193 measured models
DeepSeek V4 Pro 0813 (Non-reasoning) scores 20.4 on the Artificial Analysis Intelligence Index, ranking 242nd of 744 measured models. The closest models on this index are GPT-5.1 Codex mini (High) (20.4), Qwen3.5 Omni Plus (20.4), and o3 (20.2). At $1.98 per million tokens (3:1 input/output blended price), it is cheaper than o3 ($3.50), more expensive than GPT-5.1 Codex mini (High) ($0.688) and Qwen3.5 Omni Plus ($1.50). With an output speed of 91.2 t/s, it is faster than Qwen3.5 Omni Plus (85.3 t/s), slower than GPT-5.1 Codex mini (High) (173 t/s) and o3 (112 t/s).
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.
- GPT-5 Mini (batch)
OpenAI · — · —
20.6+0.2——Compare - GPT-5.1 Codex mini (High)
OpenAI · $0.688 · 173 t/s
20.4$0.688173 t/sCompare - Qwen3.5 Omni Plus
Alibaba · $1.50 · 85.3 t/s
20.4$1.5085.3 t/sCompare - o3
OpenAI · $3.50 · 112 t/s
20.2−0.2$3.50112 t/sCompare
