readable AI benchmarks (simplified)
DeepSeek
DeepSeek V4 Pro 0813 (Max)
DeepSeek V4 Pro 0813 (Max) has a Quality Score of 52.9, ranking 91st among 571 scored models. Per 1M tokens, pricing is $1.32 input, $3.96 output, and $0.044 cached input. It accepts text input, outputs text, and has a 1M token context window. Its Reliability Score is 50.4, ranking 122nd among 573 scored models, above average (total average: 36.5). Its Value Score is 38.7, high compared with the total Value average of 31.4.
Model specification
- Reasoning
- reasoning
- Input modalities
- text
- Output modalities
- text
- Context window
- 1M tokens
- Weights
- open weights
Token prices USD per 1M tokens
- Input
- $1.32
- Output
- $3.96
- Cached Input
- $0.044
- Cache write
- not available
- Cost per task
- $0.674
Capability
- Intelligence
- 36.0
- Coding
- not available
- Agentic
- not available
- Omniscience
- 0.8
- Correct
- 49.1%
- Blended price
- $0.691
Answer outcomes
Fully graded outcomes (Correct + Incorrect): 97.4%
Artificial Analysis benchmarks
- GDPval-AA v2
- 48%
- τ³-Banking
- 40%
- SciCode
- 51%
- Humanity’s Last Exam
- 41%
- GPQA Diamond
- 93%
- CritPt
- 18%
- AA-Omniscience
- 50%
- AA-LCR
- 80%
DeepSWE operational cross-check
- Pass@1
- 62.8%
- Avg steps/scored attempt
- 154.71
- Mean cost/scored attempt (USD)
- $0.241
- Mean output tokens/scored attempt
- 105,998.918
- Mean duration/scored attempt (s)
- 2213 s
Mean cost/scored attempt is the official feed's mean_cost_usd: mean benchmark cost in USD per scored attempt under the source's cost basis. It is not Artificial Analysis cost per task and not token price (USD per 1M tokens). Mean output tokens/scored attempt is mean_output_tokens, also averaged over every scored attempt including scored failures. Mean duration/scored attempt is mean_duration_seconds; it is informational only, and wall-clock varies with provider and host load. These three fields do not affect Quality Score, Value Score, or Reliability. Epoch General ECI's benchmark basket includes DeepSWE v1.1, so Epoch General ECI and this DeepSWE cross-check are not independent evidence.
Similar models
- DeepSeek V4.1 Flash (Max)DeepSeekQuality Score 54.0Value Score 45.9
- DeepSeek V4 Flash (high)DeepSeekQuality Score 48.2Value Score 44.2
- DeepSeek V4 Flash 0731 (Max)DeepSeekQuality Score 47.8Value Score 39.7
- DeepSeek V4 Flash Vision (Max)DeepSeekQuality Score 47.4Value Score 39.4
- DeepSeek V4 Pro (high)DeepSeekQuality Score 55.7Value Score 47.6
- DeepSeek V4 Pro 0424 (Max)DeepSeekQuality Score 45.7Value Score 39.0