DeepSeek V4 Pro deepseek-v4-pro
deepseek-ai · deepseek reasoning flagship
#7 overall
- params
- 862B
- arch
- moe
- context
- 1024k
- license
- mit open weights
- released
- Apr 2026
- reasoning
- yes
- train compute
- 9.7e+24 FLOP
- downloads/30d
- 360.4k
# architecture
- layers
- 61
- d_model
- 7168
- heads
- 128q / 1kv
- head dim
- 512
- experts
- top-6 of 384 + 1 shared
- vocab
- 129280
- family
- deepseek_v4
signal path of one layer, generated from the registry's structured fields — dimension lines quote the model's real numbers; an MoE trace forks at the router, a dense trace runs straight through. Geometry fields come from the repo's config.json.
# why ranked
Overall: #7 — score 91.5 ● high — all signals present
| signal | weight | input (0–100) |
|---|---|---|
| aa_intelligence | 0.70 | 90.2 |
| bench_composite | 0.30 | 94.4 |
benchmark panel evidence:
| complete panel | score (0–100) |
|---|---|
| aa-coding-v1 | 100.0 |
| livebench-v1 | 88.9 |
Coding: provisional — score 90.4 ● low — a single signal; treat with caution
provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.
| signal | weight | input (0–100) |
|---|---|---|
| aa_coding | 0.30 | 90.4 |
| aider_polyglot | 0.30 | — |
| swe_bench_verified | 0.40 | — |
missing signals are dropped and the remaining weights renormalized — never imputed.
# trend
methodology changed during this history window; score movement across that boundary is not model movement. See methodology v6.
# benchmarks
| benchmark | score | source | date |
|---|---|---|---|
| AA Coding Index | 68.8 | Artificial Analysis | — |
| AA Intelligence Index | 36.0 | Artificial Analysis | — |
| LiveBench Agentic Coding | 42.6 | LiveBench | Jun 2026 |
| LiveBench Coding | 70.0 | LiveBench | Jun 2026 |
| LiveBench Data Analysis | 74.5 | LiveBench | Jun 2026 |
| LiveBench IF | 62.4 | LiveBench | Jun 2026 |
| LiveBench Language | 78.1 | LiveBench | Jun 2026 |
| LiveBench Mathematics | 90.7 | LiveBench | Jun 2026 |
| LiveBench Reasoning | 82.7 | LiveBench | Jun 2026 |
# where to run
| provider | quant | ctx | $/M in | $/M out | $/M cache | price src | tps | uptime | ✓ |
|---|---|---|---|---|---|---|---|---|---|
| Featherless | — | — | — | — | via hfrouter | — | — | ||
| Nebius | — | — | — | — | nebius | — | — | ✓ | |
| DeepSeek | 1024k | $0.66 | $1.98 | $0.02 | via modelsdev | — | — | ||
| Relace | fp4 | 1024k | $0.90 | $1.50 | $0.25 | via openrouter | — | 99.6% | |
| StreamLake | fp8 | 1000k | $0.96 | $1.91 | $0.08 | via openrouter | — | 100.0% | |
| GMICloud | fp8 | 1024k | $0.96 | $1.91 | $0.08 | via openrouter | — | 99.9% | |
| DigitalOcean | unknown | 1024k | $1.04 | $2.09 | $0.21 | via openrouter | — | 100.0% | |
| Reka | unknown | 1024k | $1.05 | $2.10 | $0.21 | via openrouter | — | 99.5% | |
| Cloudflare | unknown | 1024k | $1.15 | $2.55 | $0.20 | via openrouter | — | 99.7% | |
| Wandb | 1049k | $1.15 | $2.55 | — | via litellm | — | — | ||
| DeepInfra | fp8 | 1024k | $1.30 | $2.60 | — | deepinfra | — | 99.9% | ✓ |
| Alibaba | fp8 | 1000k | $1.42 | $2.83 | $0.12 | via openrouter | — | 95.0% | |
| SiliconFlow | fp8 | 1024k | $1.50 | $3.13 | $0.14 | openrouter | — | 100.0% | ✓ |
| Alibaba | fp8 | 1000k | $1.56 | $3.12 | $0.13 | via openrouter | — | 99.0% | |
| Opencode Go | 1000k | $1.32 | $3.96 | — | via litellm | — | — | ||
| Novita | fp8 | 1024k | $1.60 | $3.20 | — | novita | — | 100.0% | ✓ |
| Venice | unknown | 1000k | $1.65 | $3.30 | — | venice | — | 98.8% | ✓ |
| AtlasCloud | fp4 | 1024k | $1.68 | $3.38 | — | atlascloud | — | 98.0% | ✓ |
| Aihubmix | 1000k | $1.69 | $3.38 | — | via litellm | — | — | ||
| Baidu | fp8 | 1024k | $1.69 | $3.38 | $0.14 | via openrouter | — | 99.4% | |
| BaseTen | 1024k | $1.74 | $3.48 | — | via litellm | — | — | ||
| Crusoe | fp8 | 1024k | $1.74 | $3.48 | $0.15 | via openrouter | — | 96.6% | |
| Tencent | 1000k | $1.74 | $3.48 | — | via litellm | — | — | ||
| Azure | unknown | 1024k | $1.91 | $3.83 | $0.16 | via openrouter | — | 100.0% | |
| Inferenceapis | 1024k | $1.72 | $5.15 | — | via litellm | — | — | ||
| Qwen Ai Platform | 1000k | $2.40 | $4.80 | — | via litellm | — | — | ||
| Qwencloud | 1000k | $2.40 | $4.80 | — | via litellm | — | — |
sorted by blended price ((3·input + output) / 4 per 1M) · ✓ = the provider's own catalog confirms the offer · "via …" prices are what the aggregator routing the offer charges, not the provider's own list price
source aliases
- aa
deepseek-v4-pro- epoch
DeepSeek-V4-Pro- hf
deepseek-ai/DeepSeek-V4-Pro-DSpark- openrouter
deepseek/deepseek-v4-pro