Best DeepSeek V3 Alternatives
DeepSeek V3 by DeepSeek is a open source model priced at $0.14/0.28 per 1M tokens (in/out). It's already affordable, but you might want different strengths or features.
DeepSeek V3
DeepSeekOpen SourceInput
$0.14/1M
Output
$0.28/1M
Context
164K
Max Output
16K
Why Switch from DeepSeek V3?
Top Alternatives
Higher benchmark scores, 128K max output, open-source and self-hostable.
Input
$0.3/1M
114% more
Output
$1.2/1M
329% more
Context
200K
Max Output
128K
Comparable performance, 1M context (6x more), adds vision.
Input
$0.31/1M
121% more
Output
$0.85/1M
204% more
Context
1M
Max Output
32K
10M context (61x more), adds vision, open-source and self-hostable.
Input
$0.18/1M
29% more
Output
$0.63/1M
125% more
Context
10M
Max Output
32K
Comparable performance, 64K max output.
Input
$0.55/1M
293% more
Output
$2.19/1M
682% more
Context
128K
Max Output
64K
Comparable performance, 128K max output, adds vision.
Input
$0.6/1M
329% more
Output
$2.2/1M
686% more
Context
200K
Max Output
128K
Comparable performance, 1M context (6x more), 66K max output.
Input
$0.15/1M
7% more
Output
$0.6/1M
114% more
Context
1M
Max Output
66K
Higher benchmark scores, 100K max output, adds vision, tool-use.
Input
$0.4/1M
186% more
Output
$1.6/1M
471% more
Context
200K
Max Output
100K
Comparable performance, 1M context (6x more), 66K max output.
Input
$0.5/1M
257% more
Output
$3/1M
971% more
Context
1M
Max Output
66K
Full Comparison Table
| Model | Input $/1M | Output $/1M | Context | MMLU-Pro | HumanEval | Score |
|---|---|---|---|---|---|---|
| MiniMax M2.5MiniMax | $0.30114% more | $1.20329% more | 200K | 82%+4.0% | 90%+1.0% | 80 |
| Llama 4 MaverickMeta | $0.31121% more | $0.85204% more | 1M | 80.5%+2.5% | 90.2%+1.2% | 75 |
| Llama 4 ScoutMeta | $0.1829% more | $0.63125% more | 10M | 74.2%-3.8% | 86%-3.0% | 68 |
| DeepSeek R1DeepSeek | $0.55293% more | $2.19682% more | 128K | 84%+6.0% | 92%+3.0% | 58 |
| GLM-4.7Zhipu AI | $0.60329% more | $2.20686% more | 200K | 84.3%+6.3% | — | 58 |
| Gemini 2.5 FlashGoogle | $0.157% more | $0.60114% more | 1M | 76%-2.0% | 89.5%+0.5% | 56 |
| o3OpenAI | $0.40186% more | $1.60471% more | 200K | 87%+9.0% | 94.5%+5.5% | 53 |
| Gemini 3 FlashGoogle | $0.50257% more | $3.00971% more | 1M | 78%Same | 90%+1.0% | 53 |
| Claude Opus 4.6Anthropic | $5.003471% more | $25.008829% more | 200K | 89.5%+11.5% | 95%+6.0% | 43 |
| GPT-5.3 CodexOpenAI | $2.001329% more | $16.005614% more | 200K | 90%+12.0% | 96.5%+7.5% | 43 |
| GPT-5.2 CodexOpenAI | $1.751150% more | $14.004900% more | 200K | 89%+11.0% | 95.5%+6.5% | 43 |
| o4-miniOpenAI | $1.10686% more | $4.401471% more | 200K | 85%+7.0% | 93.5%+4.5% | 43 |
| Mistral Large 3Mistral | $2.001329% more | $5.001686% more | 128K | 83%+5.0% | 91%+2.0% | 43 |
| GPT-4o MiniOpenAI | $0.157% more | $0.60114% more | 128K | 68%-10.0% | 87.2%-1.8% | 42 |
| GPT-5OpenAI | $1.25793% more | $10.003471% more | 128K | 88.5%+10.5% | 95%+6.0% | 40 |
| Gemini 3.1 ProGoogle | $2.001329% more | $12.004186% more | 1M | 91%+13.0% | 95%+6.0% | 40 |
| Gemini 3 ProGoogle | $2.001329% more | $12.004186% more | 1M | 89.8%+11.8% | 94%+5.0% | 40 |
| Gemini 2.5 ProGoogle | $1.25793% more | $10.003471% more | 1M | 87.5%+9.5% | 93.5%+4.5% | 40 |
| Grok 4xAI | $3.002043% more | $15.005257% more | 128K | 86%+8.0% | 93%+4.0% | 40 |
| Mistral Medium 3Mistral | $0.40186% more | $2.00614% more | 128K | 76%-2.0% | 87%-2.0% | 40 |
| GLM-5Zhipu AI | $1.00614% more | $3.201043% more | 200K | 70.4%-7.6% | 91%+2.0% | 38 |
| Claude Sonnet 4.6Anthropic | $3.002043% more | $15.005257% more | 200K | 86%+8.0% | 94%+5.0% | 36 |
| Claude Sonnet 4.5Anthropic | $3.002043% more | $15.005257% more | 200K | 84.5%+6.5% | 93%+4.0% | 36 |
| Claude Haiku 4.5Anthropic | $0.80471% more | $4.001329% more | 200K | 69.4%-8.6% | 88.1%-0.9% | 30 |
| GPT-4oOpenAI | $2.501686% more | $10.003471% more | 128K | 80.5%+2.5% | 91%+2.0% | 28 |