DeepSeek V4 Pro
DeepSeek · deepseek-ai/DeepSeek-V4-Pro
Weights license: MIT · code and components may use other terms
Quality
Quality · Overall
LiveBench 2026-06-25 · 0–100 · higher is better · this model outlined, #16 of 22
- Color = publisher
- Memory under review
Source: LiveBench 2026-06-25 · collected 2026-09-26 · open models without a score are not drawn (no score ≠ 0)
- LMArena · text · rating
- 1,458 ± 4Confidence interval 1,454–1,462 · 57,588 votes · rank 64 of the releaseLMArena 2026-10-22 · CC BY 4.0
LiveBench categories
This model's mean per category · 0–100
Source: LiveBench 2026-06-25 · collected 2026-09-26 · open models without a score are not drawn (no score ≠ 0)
Reported by the publisher
- MMLU-Pro (EM)
- 82.9V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- SimpleQA-Verified (Pass@1)
- 45.0V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- Chinese-SimpleQA (Pass@1)
- 75.8V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- GPQA Diamond (Pass@1)
- 72.9V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- HLE (Pass@1)
- 7.7V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- LiveCodeBench (Pass@1)
- 56.8V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- HMMT 2026 Feb (Pass@1)
- 31.7V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- IMOAnswerBench (Pass@1)
- 35.3V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
Every reported result
- Apex (Pass@1)
- 0.4V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- Apex Shortlist (Pass@1)
- 9.2V4-Pro Non-Think · Knowledge & Reasoning · thinking mode
- MRCR 1M (MMR)
- 44.7V4-Pro Non-Think · Long Context · thinking mode
- CorpusQA 1M (ACC)
- 35.6V4-Pro Non-Think · Long Context · thinking mode
- Terminal Bench 2.0 (Acc)
- 59.1V4-Pro Non-Think · Agentic · thinking mode
- SWE Verified (Resolved)
- 73.6V4-Pro Non-Think · Agentic · thinking mode
- SWE Pro (Resolved)
- 52.1V4-Pro Non-Think · Agentic · thinking mode
- SWE Multilingual (Resolved)
- 69.8V4-Pro Non-Think · Agentic · thinking mode
- MCPAtlas (Pass@1)
- 69.4V4-Pro Non-Think · Agentic · thinking mode
- Toolathlon (Pass@1)
- 46.3V4-Pro Non-Think · Agentic · thinking mode
- MMLU-Pro (EM)
- 87.1V4-Pro High · Knowledge & Reasoning
- SimpleQA-Verified (Pass@1)
- 46.2V4-Pro High · Knowledge & Reasoning
- Chinese-SimpleQA (Pass@1)
- 77.7V4-Pro High · Knowledge & Reasoning
- GPQA Diamond (Pass@1)
- 89.1V4-Pro High · Knowledge & Reasoning
- HLE (Pass@1)
- 34.5V4-Pro High · Knowledge & Reasoning
- LiveCodeBench (Pass@1)
- 89.8V4-Pro High · Knowledge & Reasoning
- Codeforces (Rating)
- 2,919V4-Pro High · Knowledge & Reasoning
- HMMT 2026 Feb (Pass@1)
- 94.0V4-Pro High · Knowledge & Reasoning
- IMOAnswerBench (Pass@1)
- 88.0V4-Pro High · Knowledge & Reasoning
- Apex (Pass@1)
- 27.4V4-Pro High · Knowledge & Reasoning
- Apex Shortlist (Pass@1)
- 85.5V4-Pro High · Knowledge & Reasoning
- MRCR 1M (MMR)
- 83.3V4-Pro High · Long Context
- CorpusQA 1M (ACC)
- 56.5V4-Pro High · Long Context
- Terminal Bench 2.0 (Acc)
- 63.3V4-Pro High · Agentic
- SWE Verified (Resolved)
- 79.4V4-Pro High · Agentic
- SWE Pro (Resolved)
- 54.4V4-Pro High · Agentic
- SWE Multilingual (Resolved)
- 74.1V4-Pro High · Agentic
- BrowseComp (Pass@1)
- 80.4V4-Pro High · Agentic
- HLE w/ tools (Pass@1)
- 44.7V4-Pro High · Agentic
- MCPAtlas (Pass@1)
- 74.2V4-Pro High · Agentic
- Toolathlon (Pass@1)
- 49.0V4-Pro High · Agentic
- MMLU-Pro (EM)
- 87.5V4-Pro Max · Knowledge & Reasoning
- SimpleQA-Verified (Pass@1)
- 57.9V4-Pro Max · Knowledge & Reasoning
- Chinese-SimpleQA (Pass@1)
- 84.4V4-Pro Max · Knowledge & Reasoning
- GPQA Diamond (Pass@1)
- 90.1V4-Pro Max · Knowledge & Reasoning
- HLE (Pass@1)
- 37.7V4-Pro Max · Knowledge & Reasoning
- LiveCodeBench (Pass@1)
- 93.5V4-Pro Max · Knowledge & Reasoning
- Codeforces (Rating)
- 3,206V4-Pro Max · Knowledge & Reasoning
- HMMT 2026 Feb (Pass@1)
- 95.2V4-Pro Max · Knowledge & Reasoning
- IMOAnswerBench (Pass@1)
- 89.8V4-Pro Max · Knowledge & Reasoning
- Apex (Pass@1)
- 38.3V4-Pro Max · Knowledge & Reasoning
- Apex Shortlist (Pass@1)
- 90.2V4-Pro Max · Knowledge & Reasoning
- MRCR 1M (MMR)
- 83.5V4-Pro Max · Long Context
- CorpusQA 1M (ACC)
- 62.0V4-Pro Max · Long Context
- Terminal Bench 2.0 (Acc)
- 67.9V4-Pro Max · Agentic
- SWE Verified (Resolved)
- 80.6V4-Pro Max · Agentic
- SWE Pro (Resolved)
- 55.4V4-Pro Max · Agentic
- SWE Multilingual (Resolved)
- 76.2V4-Pro Max · Agentic
- BrowseComp (Pass@1)
- 83.4V4-Pro Max · Agentic
- HLE w/ tools (Pass@1)
- 48.2V4-Pro Max · Agentic
- MCPAtlas (Pass@1)
- 73.6V4-Pro Max · Agentic
- GDPval-AA (Elo)
- 1,554V4-Pro Max · Agentic
- Toolathlon (Pass@1)
- 51.8V4-Pro Max · Agentic
Across DeepSeek V4 sizes
| Benchmark (publisher settings) | Flash 284B | Pro 1.6T |
|---|---|---|
| MMLU-Pro (EM) | 83.0 | 82.9 |
| SimpleQA-Verified (Pass@1) | 23.1 | 45.0 |
| Chinese-SimpleQA (Pass@1) | 71.5 | 75.8 |
| GPQA Diamond (Pass@1) | 71.2 | 72.9 |
| HLE (Pass@1) | 8.1 | 7.7 |
| LiveCodeBench (Pass@1) | 55.2 | 56.8 |
| HMMT 2026 Feb (Pass@1) | 40.8 | 31.7 |
| IMOAnswerBench (Pass@1) | 41.9 | 35.3 |
| Apex (Pass@1) | 1.0 | 0.4 |
| Apex Shortlist (Pass@1) | 9.3 | 9.2 |
| MRCR 1M (MMR) | 37.5 | 44.7 |
| CorpusQA 1M (ACC) | 15.5 | 35.6 |
| Terminal Bench 2.0 (Acc) | 49.1 | 59.1 |
| SWE Verified (Resolved) | 73.7 | 73.6 |
| SWE Pro (Resolved) | 49.1 | 52.1 |
| SWE Multilingual (Resolved) | 69.7 | 69.8 |
| MCPAtlas (Pass@1) | 64.0 | 69.4 |
| Toolathlon (Pass@1) | 40.7 | 46.3 |
| MMLU-Pro (EM) | 86.4 | 87.1 |
| SimpleQA-Verified (Pass@1) | 28.9 | 46.2 |
| Chinese-SimpleQA (Pass@1) | 73.2 | 77.7 |
| GPQA Diamond (Pass@1) | 87.4 | 89.1 |
| HLE (Pass@1) | 29.4 | 34.5 |
| LiveCodeBench (Pass@1) | 88.4 | 89.8 |
| Codeforces (Rating) | 2,816 | 2,919 |
| HMMT 2026 Feb (Pass@1) | 91.9 | 94.0 |
| IMOAnswerBench (Pass@1) | 85.1 | 88.0 |
| Apex (Pass@1) | 19.1 | 27.4 |
| Apex Shortlist (Pass@1) | 72.1 | 85.5 |
| MRCR 1M (MMR) | 76.9 | 83.3 |
| CorpusQA 1M (ACC) | 59.3 | 56.5 |
| Terminal Bench 2.0 (Acc) | 56.6 | 63.3 |
| SWE Verified (Resolved) | 78.6 | 79.4 |
| SWE Pro (Resolved) | 52.3 | 54.4 |
| SWE Multilingual (Resolved) | 70.2 | 74.1 |
| BrowseComp (Pass@1) | 53.5 | 80.4 |
| HLE w/ tools (Pass@1) | 40.3 | 44.7 |
| MCPAtlas (Pass@1) | 67.4 | 74.2 |
| Toolathlon (Pass@1) | 43.5 | 49.0 |
| GDPval-AA (Elo) | 1,395 | 1,554 |
Values as printed on the card, with the publisher's own settings (card section: Evaluation Results › Instruct Model › Comparison across Modes). Not comparable across publishers and not used in any ranking. Model card
Quality × memory
Quality × memory
LiveBench 2026-06-25 overall vs. Q4 planning memory (log scale) · 18 open models
- This model
- Color = publisher
Memory = the model's own figure: parameters × 0.625 bytes + 20% working room, before any machine's reserve; with a chosen machine, the total there with its reserve (25% of memory, 2 to 8 GB; 2 GB per graphics card). An estimate, not measured VRAM. LiveBench 2026-06-25
Runs on
More memory than any profile in the catalog
Even the largest profile (8 × 80 GB, aggregated) is below this planning estimate. Multi-node setups are outside this catalog.
DeepSeek V4 Pro is a text model from DeepSeek with 1,600B parameters, 49B of them active per token (mixture of experts). It was released on April 24, 2026 and reads up to 1,048,576 tokens of context. Its Q4 weights plus working room come to about 1,200 GB, more than the fast memory of every reference machine. No GGUF or MLX file for it was collected from the Hub, so this page lists no file sizes or run commands. It scores 71.6 overall in LiveBench 2026-06-25. Its LMArena text rating is 1,458 ± 4 (release 2026-10-22), a separate human-preference scale.
On the reference machines
| Where Q4 sits | ||
|---|---|---|
| Mac mini M2 · 8 GB unifiedLarger than its memory | ~1,202 GB | Larger than its memory |
| Mac mini M4 · 16 GB unifiedLarger than its memory | ~1,204 GB | Larger than its memory |
| Mac mini M4 · 24 GB unifiedLarger than its memory | ~1,206 GB | Larger than its memory |
| Mac mini M4 · 32 GB unifiedLarger than its memory | ~1,208 GB | Larger than its memory |
| Mac mini M5 Pro · 64 GB unifiedLarger than its memory | ~1,208 GB | Larger than its memory |
| Mac Studio M5 Max · 128 GB unifiedLarger than its memory | ~1,208 GB | Larger than its memory |
| Mac Studio · M5 Ultra · 36-core CPU / 80-core GPU · 512 GB unifiedLarger than its memory | ~1,208 GB | Larger than its memory |
| RTX 4060 · 8 GB VRAMLarger than its memory | ~1,202 GB | Larger than its memory |
| RTX 3060 · 12 GB VRAMLarger than its memory | ~1,202 GB | Larger than its memory |
| RTX 5080 · 16 GB VRAMLarger than its memory | ~1,202 GB | Larger than its memory |
| RTX 4090 · 24 GB VRAMLarger than its memory | ~1,202 GB | Larger than its memory |
| RTX 5090 · 32 GB VRAMLarger than its memory | ~1,202 GB | Larger than its memory |
| DGX H100 · 8 × 80 GBLarger than its memory | ~1,216 GB | Larger than its memory |
No GGUF or MLX file was collected for this checkpoint, so the table has no file column; the Q4 figures are planning estimates from the parameter count, not measured runs.
Compare DeepSeek V4 Pro using its publisher card and the LiveBench 2026-06-25 snapshot. This is quality evidence, not local speed.
The Q4 weights plus working room come to about 1200 GB from 1,600B stated parameters, before the machine's reserve (Runs on adds it per machine). It is not a measured minimum; context, batch size, runtime and quantization change actual memory use.
Measured speed
No published measurement for this checkpoint
Tokens per second are shown only when a source publishes them with hardware, build and setup. They are never estimated.
How to run
The memory needed above is the Q4 load before a machine's reserve. Formula: parameters × 0.625 bytes, plus 20% working allowance; each machine then adds its system reserve: 25% of unified or system memory, 2 to 8 GB, or 2 GB per GPU on dedicated cards. This is neither measured VRAM nor checkpoint download size.
Q4 planning estimate only. Check checkpoint format, runtime support, KV cache, context, batch size and offload before choosing hardware.
Documented ways to run it
The model card points to another page without a command of its own
Evidence level: Generic estimate only; no execution reproduced by this site.
- Versions to download
- — No GGUF or MLX file collected from the Hub for this checkpoint
- Publisher sampling defaults
- — No Hub file facts collected for this checkpoint
- Chat template markers
- — Chat template not read for this checkpoint
Before downloading
- Check the exact checkpoint and license linked here.
- Confirm a compatible runtime and precision or quantization for your platform.
- Allow for context, batch size and any extra media components beyond the weights.
Technical details
What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.
- Checkpoint / version
- deepseek-ai/DeepSeek-V4-Pro
- Task and modality
- language and reasoning · text
- Count and basis
- 1,600B · manufacturer specification · 49B active
- Context
- 1,048,576 tokens published for this model
- Architecture
- MoE
- Released
- Change Log | DeepSeek API Docs · read 2026-10-09source
- Runtime and precision
- No local execution verified.
- Commercial use
- Allowed MIT license text (SPDX) · read 2026-10-18license text
Every stored field
- Revision checked
- b5968e9190ef611bbf34a7229255be88a0e937c1
- Architecture
- DeepseekV4ForCausalLM
- Library
- transformers
- Hub task tag
- text-generation
- Formats
- safetensors
- Access
- Open download
- Publisher kind
- Official
- Repository created
- 2026-04-22creation is not the model's release
- Repository updated
- 2026-06-22a commit is not a release date
- Hub downloads (30 days) · likes
- 360,375 · 5,677this repository only · read 2026-10-25
- Catalog state
- Hub config or weights found; no inventory benchmark link
Sources
Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.
- Publisher checkpoint and model card
- Weight license and use terms
- Quality benchmark source · LiveBench 2026-06-25 overall
Catalog reviewed: 2026-09-30. Benchmark release: 2026-06-25; snapshot collected: 2026-09-26.