DeepSeek-V3
Model ID
- Context
- 65.54K
- 8K max output
- Input price
- $0.27
- per 1M tokens · DeepSeek API
- Output price
- $1.10
- per 1M tokens
- Epoch Capabilities Index
- 132.3
- #100 of 105 listed
- LMArena Text
- 1,358
- #121 of 149 listed
Where to run it
4 offerings · prices as listed| Meter | Tier | Price |
|---|---|---|
DeepSeek API (opens in a new tab) deepseek-v3 | ||
| Input | Standard | $0.27 / 1M tokens |
| Output | Standard | $1.10 / 1M tokens |
| Cache read | Standard | $0.070 / 1M tokens |
Alibaba Cloud Model Studio (opens in a new tab) deepseek-v3 | ||
| Input | Standard | $0.29 / 1M tokens |
| Output | Standard | $1.15 / 1M tokens |
Alibaba Cloud Model Studio (opens in a new tab) deepseek-v3 | ||
| Input | Standard | ¥2.00 / 1M tokens |
| Output | Standard | ¥8.00 / 1M tokens |
Alibaba Cloud Model Studio (opens in a new tab) vanchin/deepseek-v3 | ||
| Input | Standard | ¥2.00 / 1M tokens |
| Output | Standard | ¥8.00 / 1M tokens |
Changes
All changes →- Now priced on DeepSeek API Input $0.27 · Output $1.10 · Cache read $0.070
- Now priced on Alibaba Cloud Model Studio, China Input ¥2.00 · Output ¥8.00
- Context window 128K → 65.54K
Cited scores
From third-party leaderboards, not measured by us- Epoch Capabilities Index 132.3
#100 of 105 listed models · CI 127.5–135.5
- LMArena Text 1,358
#121 of 149 listed models · CI 1,354–1,363 · 21,770 votes
- Epoch AI · GPQA Diamond 56.5%
#96 of 128 listed models · ±2.76
- LMArena Text · Chinese 1,369
#123 of 149 listed models · CI 1,354–1,385 · 1,450 votes
- LMArena Text · Coding 1,387
#128 of 149 listed models · CI 1,377–1,397 · 3,280 votes
- Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Oct 3, 2026
- LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Oct 2, 2026
- changes we made