Overview
Deepseek
Models
24
Cheapest Input
$0.01
Max Context
1.0M
Top Intelligence
39.3
Meta-llama
Models
42
Cheapest Input
$0.02
Max Context
1.0M
Top Intelligence
18.4
Flagship Model Comparison
DeepSeek V4 ProvsLlama 4 Maverick
| Metric | DeepSeek V4 Pro | Llama 4 Maverick |
|---|---|---|
| Input $/1M tokens | $0.43 | $0.15 |
| Output $/1M tokens | $0.87 | $0.60 |
| Context Window | 1.0M | 1.0M |
| Intelligence | 39.3 | 18.4 |
| Coding | 38.4 | 15.6 |
| Math | N/A | 19.3 |
All Models Pricing
Deepseek Models
| Model | Input/1M | Output/1M | Context |
|---|---|---|---|
| DeepSeek V4 Pro | $0.43 | $0.87 | 1.0M |
| DeepSeek V4 Flash (Non-Reasoning) | $0.14 | $0.28 | 1.0M |
| DeepSeek V3.2 | $0.25 | $0.38 | 164K |
| DeepSeek V3.2 Speciale | $0.29 | $0.43 | 164K |
| DeepSeek V3.1 Terminus | $0.27 | $0.95 | 164K |
| DeepSeek V3.2 Exp | $0.27 | $0.41 | 164K |
| DeepSeek V3.1 | $0.15 | $0.75 | 33K |
| R1 0528 | $0.50 | $2.15 | 164K |
| R1 | $0.55 | $2.00 | 64K |
| R1 Distill Qwen 32B | $0.29 | $0.29 | 33K |
| DeepSeek V3 0324 | $0.20 | $0.77 | 164K |
| DeepSeek R1 0528 Qwen3 8B | $0.20 | $0.20 | 128K |
| R1 Distill Llama 70B | $0.70 | $0.80 | 131K |
| R1 Distill Qwen 14B | $0.15 | $0.15 | 33K |
| R1 Distill Llama 8B | $0.04 | $0.04 | 33K |
| R1 Distill Qwen 1.5B | $0.18 | $0.18 | 131K |
| DeepSeek V3 | $0.01 | $0.03 | 164K |
| DeepSeek Coder 1.3B Base | $0.10 | $0.10 | 16K |
| DeepSeek Coder 7B Base | $0.20 | $0.20 | 4K |
| R1 Distill Qwen 7B | $0.20 | $0.20 | 33K |
| DeepSeek Coder 7B Base v1.5 | $0.20 | $0.20 | 4K |
| DeepSeek Coder 7B Instruct v1.5 | $0.20 | $0.20 | 4K |
| DeepSeek Prover V2 | $0.50 | $2.18 | 164K |
| DeepSeek Coder 33B Instruct | $0.80 | $0.80 | 16K |
Meta-llama Models
Benchmark Comparison
Best Scores by Provider
| Benchmark | Deepseek | Meta-llama |
|---|---|---|
| Intelligence | 39.3 DeepSeek V4 Pro | 18.4 Llama 4 Maverick |
| Coding | 38.4 DeepSeek V4 Pro | 15.6 Llama 4 Maverick |
| Math | 96.7 DeepSeek V3.2 Speciale | 19.3 Llama 4 Maverick |
| MMLU Pro | 86.3 DeepSeek V3.2 Speciale | 80.9 Llama 4 Maverick |
| GPQA | 87.1 DeepSeek V3.2 Speciale | 67.1 Llama 4 Maverick |
| LiveCodeBench | 89.6 DeepSeek V3.2 Speciale | 39.7 Llama 4 Maverick |
| Aider | 74.2 DeepSeek V3.2 Exp | 66.2 Llama 3.1 405B Instruct |
| AIME | 89.3 R1 0528 | 39.0 Llama 4 Maverick |
Capabilities
| Capability | Deepseek | Meta-llama |
|---|---|---|
| Vision | — | ✓ (4 models) |
| Tool Calls | ✓ (18 models) | ✓ (14 models) |
| Reasoning | ✓ (15 models) | — |
| Audio Input | — | — |
| Audio Output | — | — |
| PDF Input | — | — |
| Web Search | — | — |
| Prompt Caching | ✓ (6 models) | — |
| Open Source Models | ✓ (18 models) | ✓ (42 models) |
Model-Level Comparisons
Compare specific models head-to-head: