Compare models
Pick up to 4models. Specs render side-by-side. Share the URL — it's stateless.
Comparing agents instead? Switch to agent compare →
 DeepSeek-V4-ProDeepSeek | DeepSeek-R1 DeepSeek |  Qwen3.7-MaxAlibaba | |
|---|---|---|---|
| Vendor | DeepSeek | DeepSeek | Alibaba |
| Family | DeepSeek | DeepSeek | Qwen |
| Release date | 2026-04-22 | 2025-01-20 | 2026-05-20 |
| Context window | 1,000,000 tokens | 128,000 tokens | 1,000,000 tokens |
| Parameters | 1.6T (49B active) | 671B | — |
| Modality | text | text | text |
| License | MIT | MIT | proprietary |
| Source | open weights | open weights | proprietary |
| Description | DeepSeek's flagship open-weight MoE. 1.6T parameters with 49B activated, 1M-token context, and a hybrid attention scheme (CSA + HCA) that delivers long-context inference at ~27% of V3.2's FLOPs. | Reasoning-focused open-weight model. | Alibaba's flagship agent model — 1M-token context, extended-thinking mode, 56.6 on the Artificial Analysis Intelligence Index v4.0 (5th overall, #1 Chinese). 50.8% on Terminal-Bench Hard. Designed for long-horizon agent workloads (hundreds-to-thousands of steps). Closed-weight, $2.50/$7.50 per 1M tokens. |
| Links | |||
| Benchmarks | |||
| MMLU-Pro | 84.2% | — | 83.7% |
| GPQA-D | 82.4% | — | 83.0% |
| HumanEval | 95.1% | — | 93.9% |
| Aider | 80.1% | — | 78.4% |
| AIME-25 | 88.6% | 79.8% | 90.4% |
| LiveCB | 72.4% | — | 71.0% |
Add a model
Claude 4.8 Opus+Qwen3.7-Max✓
Gemini 3.5+Gemini Omni+
GPT-5.5+Qwen3.6-27B+DeepSeek-V4-Flash+DeepSeek-V4-Pro✓Gemini 3 Pro+Claude 4.7 Opus+Claude 4.6 Sonnet+Claude 4.5 Haiku+DeepSeek-V3.1+GPT-5+
Grok 4+Gemini 2.5 Flash+Qwen3 235B+o3+GPT-4.1+
Llama 4 Scout+Llama 4 Maverick+Gemini 2.5 Pro+DeepSeek-R1✓
Mistral Large 2+