Buying guide
| Input | $4.00 | $0.300 |
| Output | $20.00 | $1.20 |
| Cached input | $0.400 | $0.0060 |
| Context window | 1,050K | 1,000K |
| Blended /1M | $12.00 | $0.750 |
DeepSeek V4.1-Flash 0910 is the cheaper option per blended million tokens.
You need the highest benchmark scores and can absorb the higher per-token price.
You are optimizing for cost at scale and your workloads tolerate the score gap.
GPT 5.6-Sol leads on 9 of 16 shared benchmarks; DeepSeek V4.1-Flash 0910 is ~94% cheaper per blended million tokens.
| Dimension | GPT 5.6-Sol | DeepSeek V4.1-Flash 0910 |
|---|---|---|
| GPQA Diamond | 94.1% | 90.9% |
| HLE | 44.5% | 36.8% |
| Codeforces (Rating) | — | 3471 |
| MathArena Apex | — | 65.6% |
| Terminal-Bench 2.1 | 88.8% | 90.6% |
| Terminal-Bench 3.0 | 34.4% | 30% |
| Terminal-Bench 4.0 | 39.9% | 31.2% |
| DeepSWE v1.1 | 73% | 74.2% |
| ProgramBench | 23% | 20.3% |
| NL2Repo-Bench | 56.8% | 65.4% |
| CyberGym | 84.5% | 88.1% |
| SEC-Bench Pro | 74.3% | 62.8% |
| ExploitGym | 33.7% | 15.3% |
| HLE (w/tools) | — | 63.9% |
| Automation-Bench | 45.8% | 54.8% |
| Agents' Last Exam | 26.7% | 31.8% |
| Chartography (w/tools) | 79.9% | 78.9% |
| BabyVision (w/tools) | 88.9% | 89.6% |
| ZeroBench-main (w/tools) | 53% | 49% |
| Blended price | $12.00 /1M | $0.750 /1M |
| Full priceInput · Output · Cached input | $4.00 · $20.00 · $0.400 | $0.300 · $1.20 · $0.0060 |
| Context window | 1,050K | 1,000K |
Adjust the monthly volume — every model recomputes instantly.
Estimates use list prices and ignore cached input, batch discounts, and off-peak rates.