DeepSeek Flash vs gpt-6.1-terra
DeepSeek Flash is 3.8× cheaper: $0.525 per million tokens blended, against $2.00 for gpt-6.1-terra. On the Vals Index v2.2, DeepSeek Flash scores 61.26% and gpt-6.1-terra 60.54%, so DeepSeek Flash leads.
Price
| Per million tokens | DeepSeek Flash | gpt-6.1-terra |
|---|---|---|
| Maker | DeepSeek | OpenAI |
| Input | $0.3 | $1.00 |
| Output | $1.20 | $5.00 |
| Cache read | $0.006 | $0.05 |
| Blended (3:1) | $0.525 | $2.00 |
| Context window | n/a | 1,050K |
| Output speed | 177 tokens/s | 91 tokens/s |
| Status | Available | Available |
Sources
DeepSeek: official API pricing, read 2026-11-12
OpenAI: official API pricing, read 2026-11-20
Benchmarks
| Benchmark | DeepSeek Flash | gpt-6.1-terra |
|---|---|---|
| Vals Index v2.2 Vals AI | 61.26%V4.1 Flash 1017 | 60.54% |
| Artificial Analysis Intelligence Index Artificial Analysis | 33V4.1 Flash 1017, max | 51max |
| Epoch Capabilities Index Epoch AI | 159.4V4.1 Flash 1017 | 167.7 |
| Arena Text Arena | 1475V4.1 Flash 0910, Max | 1474Max |
| Terminal-Bench 4.0 Vals AI | 31.82%V4.1 Flash 1017 | 49.49% |