DeepSeek
DeepSeek V4 Pro
$1.63 / 1M tokens
cheapest at DeepInfra · $1.30 in · $2.60 out
Wins 3 of 4
Head to head · prices updated Oct 31, 2026
GPT-6.1 Sol costs 2.5× as much as DeepSeek V4 Pro at their cheapest providers (blended 3 input : 1 output tokens).
DeepSeek
$1.63 / 1M tokens
cheapest at DeepInfra · $1.30 in · $2.60 out
Wins 3 of 4
OpenAI
$4.00 / 1M tokens
cheapest at OpenAI · $2.00 in · $10.00 out
Wins 1 of 4
Verdict
GPT-6.1 Sol costs 2.5× as much as DeepSeek V4 Pro at their cheapest providers (blended 3 input : 1 output tokens).
GPT-6.1 Sol scores higher on GPQA Diamond: 95.4% (max reasoning) vs 90.9% (high reasoning).
So the choice is a trade-off: GPT-6.1 Sol for the hardest reasoning, DeepSeek V4 Pro when cost matters more than the last few points of accuracy.
DeepSeek V4 Pro accepts longer prompts (1049K vs 1000K tokens).
Your monthly bill
DeepSeek V4 Pro saves you $21.80 a month ($261.60 a year) for this workload.
| DeepSeek V4 Pro | GPT-6.1 Sol | |
|---|---|---|
| Cheapest blended price / 1M | $1.63 | $4.00 |
| Input / output per 1M | $1.30 / $2.60 | $2.00 / $10.00 |
| Accuracy (GPQA Diamond) | 90.9% (high reasoning) | 95.4% (max reasoning) |
| Max context | 1M | 1M |
| Providers we track | 6 | 3 |
Standard list prices at each model's cheapest provider we track; blended = 3 input : 1 output tokens. GPQA Diamond from Epoch AI (CC BY 4.0), best measured reasoning setting, named next to each score. Methodology.
GPT-6.1 Sol costs 2.5× as much as DeepSeek V4 Pro at their cheapest providers (blended 3 input : 1 output tokens).
On GPQA Diamond (graduate-level science questions, measured by Epoch AI) DeepSeek V4 Pro scores 90.9% (high reasoning) and GPT-6.1 Sol 95.4% (max reasoning). One benchmark is not the whole picture; test both on your own prompts.
From $1.30 input and $2.60 output per 1M tokens at DeepInfra, the cheapest of 6 providers we track.
From $2.00 input and $10.00 output per 1M tokens at OpenAI, the cheapest of 3 providers we track.
Swap a model