Work Automations
Private 657 tasks; 50 turns; objective completion
- 01
- 02
- 03
- 04
- 05
- 07
DeepSeek V4.1 Flash
Across the task benchmarks we track, DeepSeek V4.1 Flash’s strongest relative results are in work automations and knowledge work. On AutomationBench-AA, it scores 68.89%, ranking #7 among 55 models with reported results in our dataset.
Updated: Data as of:
Strongest results
| Benchmark | Measures | Score |
|---|---|---|
| AutomationBench-AA ↗1.0.6 / AA private 657 | Work automations | 68.89%#7 of 55 |
| AA-Briefcase ↗1.1 / combined Elo | Business deliverables | 1,416.94 Elo#22 of 57 |
| Terminal-Bench ↗4.0 / mini-swe-agent | Working in a terminal | 26.77%#22 of 56 |
| GDPval-AA ↗2.1 | Knowledge work | 1,600 Elo#16 of 38 |
Model details
| Model | DeepSeek V4.1 Flash |
|---|---|
| Agent Index v1.2 | Insufficient coverage (4/7) · Preview |
| Creator | DeepSeek (opens in a new tab) |
| Availability | Open source / open weights |
| Context size | 1,000,000 tokens (opens in a new tab) |
| Max output | Not reported (opens in a new tab) |
| Knowledge cutoff | Not reported |
| License | MIT · Model weights ↗ (opens in a new tab) |
Pricing and performance
Per provider and reported configuration. Prices are USD per 1M tokens.
| Provider / configuration | Contexttokens | InputUSD / 1M tokens | OutputUSD / 1M tokens | LatencyTTFT · seconds | Speedtokens / sec |
|---|---|---|---|---|---|
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 0.63 (opens in a new tab) | 244.981 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) (FAST) | 1,048,576 | $0.60 (opens in a new tab) | $2.40 (opens in a new tab) | 0.94 (opens in a new tab) | 830.034 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.144 (opens in a new tab) | $0.576 (opens in a new tab) | 1.515 (opens in a new tab) | 289.729 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.20 (opens in a new tab) | $0.65 (opens in a new tab) | 0.997 (opens in a new tab) | 297.861 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,048,576 | $0.29 (opens in a new tab) | $1.20 (opens in a new tab) | 0.771 (opens in a new tab) | 205.633 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 0.91 (opens in a new tab) | 384.067 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.14 (opens in a new tab) | $0.42 (opens in a new tab) | 0.698 (opens in a new tab) | 164.074 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 0.821 (opens in a new tab) | 81.897 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.18 (opens in a new tab) | $0.72 (opens in a new tab) | 1.813 (opens in a new tab) | 140.702 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,048,576 | $0.26 (opens in a new tab) | $1.17 (opens in a new tab) | 0.662 (opens in a new tab) | 190.722 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 0.559 (opens in a new tab) | 157.789 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,048,576 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 1.334 (opens in a new tab) | 257.082 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 524,288 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 1.228 (opens in a new tab) | 136.759 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,048,576 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 0.63 (opens in a new tab) | 239.17 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 262,144 | $0.70 (opens in a new tab) | $2.09 (opens in a new tab) | 1.41 (opens in a new tab) | 201.347 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,000,000 | $0.15 (opens in a new tab) | $0.60 (opens in a new tab) | 1.596 (opens in a new tab) | 154.517 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,048,576 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 0.311 (opens in a new tab) | 373.034 (opens in a new tab) |
| DeepSeek V4.1 Flash (max) | 1,048,576 | $0.30 (opens in a new tab) | $1.20 (opens in a new tab) | 0.563 (opens in a new tab) | 320.504 (opens in a new tab) |
Benchmark performance
DeepSeek V4.1 Flash in color, alongside leading models. Latest stored reports, best effort per report date.
AutomationBench-AA ↗
Knowledge Work
220 tasks; Stirrup agent; pairwise Elo
- 01
- 02
- 03
- 04
- 05
- 16
Reported benchmarks
All published results in our dataset, with versions, evaluation settings, and sources.
| Benchmark | Score | Evaluation setup | Source |
|---|---|---|---|
| AA-Briefcase1.1 / combined EloCombined Elo · Higher is better | 1,416.94 Elo | selection: best reported configuration per release · DeepSeek V4.1 Flash (Max) · tools: benchmark-defined | Artificial Analysis (opens in a new tab)independent |
| AutomationBench-AA1.0.6 / AA private 657Objectives completed without guardrail violations · Higher is better | 68.891% | DeepSeek V4.1 Flash (Max); Private 657 tasks; 50 turns; objective completion | Artificial Analysis (opens in a new tab)independent |
| AutomationBench-AA1.0.6 / AA private 657Objectives completed without guardrail violations · Higher is better | 47.808% | DeepSeek V4.1 Flash (Non-reasoning); Private 657 tasks; 50 turns; objective completion | Artificial Analysis (opens in a new tab)independent |
| GDP.pdf100 tasks / All-pass / five attemptsAccuracy · Higher is better | 12.8% | selection: best reported configuration per release · DeepSeek V4.1 Flash (Max) · tools: benchmark-defined | Artificial Analysis (opens in a new tab)independent |
| GDPval-AA2.1Elo · Higher is better | 1,600 Elo | DeepSeek V4.1 Flash (Max); 220 tasks; Stirrup agent; pairwise Elo | Artificial Analysis (opens in a new tab)independent |
| GDPval-AA2.1Elo · Higher is better | 1,348.71 Elo | DeepSeek V4.1 Flash (Non-reasoning); 220 tasks; Stirrup agent; pairwise Elo | Artificial Analysis (opens in a new tab)independent |
| HLE-Diamond1,000 questions / no toolsAccuracy · Higher is better | 19.1% | selection: best reported configuration per release · DeepSeek V4.1 Flash (Max) · No tools | Artificial Analysis (opens in a new tab)independent |
| Humanity's Last ExamHLE — source does not specify subsetAccuracy · Higher is better | 63.9% | subset: unspecified · Tools enabled | DeepSeek ↗ (opens in a new tab)provider reported · 2026-09-10 |
| Humanity's Last ExamStable HLE — full setAccuracy · Higher is better | 36.8% | subset: full · No tools | DeepSeek ↗ (opens in a new tab)provider reported · 2026-09-10 |
| Humanity's Last ExamStable HLE — text-only subsetAccuracy · Higher is better | 39.1% | subset: text · No tools | DeepSeek ↗ (opens in a new tab)provider reported · 2026-09-10 |
| Humanity's Last ExamStable HLE — text-only subsetAccuracy · Higher is better | 39.2% | question count: 2158 · reported precision decimals: 1 · DeepSeek V4.1 Flash (max); original HLE; 2,158 text-only questions; no tools; pass@1; chart labels rounded to one decimal. · source model href: /models/deepseek-v4-1-flash · subset: text · No tools | Artificial Analysis (opens in a new tab)independent |
| Terminal-Bench4.0 / mini-swe-agentTask success · Higher is better | 26.768% | harness: mini-swe-agent · repeats: 3 · selection: best reported configuration per release · DeepSeek V4.1 Flash (Max) | Artificial Analysis (opens in a new tab)independent |