DeepSeek V3
by DeepSeek · deepseek/deepseek-chat · #154 cheapest of 422 paid models
Prices updated Sep 19, 2026, 6:21 PM UTC · refreshed hourly
DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions. Pre-trained on nearly 15 trillion tokens, the reported evaluations...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$0.32
per 1M tokens
Output
$0.89
per 1M tokens
Blended (3:1)
$0.462
3 input : 1 output
Cache read
—
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for DeepSeek V3 (no cache-read rate published — hits billed as normal input).
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $0.462 | $0.0335 | — |
| 50% | $0.462 | $0.0335 | −0% |
| 90% | $0.462 | $0.0335 | −0% |
*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates (available for OpenAI / Anthropic / Google — not flagged for this creator). Confirm on the vendor’s pricing page.
What a request costs
| Workload | Input tokens | Output tokens | Cost / request | Cost / 1K requests |
|---|---|---|---|---|
| Short chat message | 500 | 300 | $0.000427 | $0.427 |
| Document summary | 8,000 | 1,000 | $0.00345 | $3.45 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.0114 | $11.38 |
| Long-context analysis | 150,000 | 5,000 | $0.0525 | $52.45 |
Performance
Intelligence Index
8.5
AA composite quality
Coding Index
23
Math Index
26
Agentic Index
—
Output speed
0 tok/s
median
Time to first token
0s
median
Time to first answer
0s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| MMLU-Pro | 75.2% |
| GPQA Diamond | 55.7% |
| Humanity's Last Exam | 2.9% |
| LiveCodeBench | 35.9% |
| SciCode | 35.8% |
| MATH-500 | 88.7% |
| AIME | 25.3% |
| AIME 2025 | 26.0% |
| IFBench | 34.8% |
| AA-LCR (Long Context Reasoning) | 29.3% |
| Terminal-Bench Hard | 6.8% |
| Terminal-Bench 2.1 | 16.9% |
| τ²-Bench (Telecom) | 22.8% |
| τ³-Bench Banking | 4.7% |
Design Arena Elo
| Category | Elo |
|---|---|
| Website | 1,132 |
| UI components | 1,108 |
| Data viz | 1,103 |
| SVG | 1,000 |
| 3D | 1,115 |
| Game dev | 1,077 |
| Code categories | 1,123 |
Specs
Context window
164K
tokens
Max output
16K
tokens
Input modalities
text
Tokenizer
DeepSeek