- Score
- 8.6 (+0.33 points)
- Total Cost (Current Price)
- $0.061 · 92.0% cheaper
- Response Time (avg)
- 7.51s · 90.4% faster
DeepSeek
Release: 2026-10-18
Tested on: 2026-10-18 00:18
deepseek/deepseek-v4.1-flash-1017::max
~748B total (~16B active)MoEOpen sourceEstimated
Summary
DeepSeek V4.1 Flash 1017 scores 8.2 on AI BENCHY and ranks #100. It has 10.0 reliability, a 76.8% pass rate, $0.757 total cost, and 78.20s average response time.
What makes DeepSeek V4.1 Flash 1017 unique: It stands out most in Domain specific, where it ranks #2, while Coding is its weakest area at #17. It uses unusually many reasoning tokens, which can help explain its slower or more expensive runs.
Recommended alternatives
We recommend choosing one of these models instead:
- Score
- 8.3 (+0.02 points)
- Total Cost (Current Price)
- $0.047 · 93.8% cheaper
- Response Time (avg)
- 18.07s · 76.9% faster
- Score
- 9.4 (+1.12 points)
- Total Cost (Current Price)
- $0.183 · 75.9% cheaper
- Response Time (avg)
- 6.24s · 92.0% faster
Model facts
Researched on 2026-10-18
- Parameters
- ~748B total (~16B active)
- Architecture
- MoE
- Availability
- Open source
- License
- MIT
Best estimate from public evidence; the vendor did not disclose every value. The exact 1017 checkpoint publishes MIT weights, prompt encoding and runnable inference code. Its card describes a minor update to V4.1 Flash and links the family technical report. Size estimates use the family 552B backbone plus 196B Engram memory and 16B active decode (8B prefill), supported by the exact checkpoint MoE/Engram configuration; the 1017 card does not separately restate counts. General multimodal LLM, not a vision specialist. OpenRouter declares none, low, high and max reasoning modes. Seven current providers publish no-training endpoint policies; routing uses existing OpenRouter account guardrails as requested. Fallback release date is the onboarding request date.
8.2
Consistency
8.2
10.0
$0.757
Total Output Tokens
1,170,181
Total Input Tokens
360,065
Input Price
$0.150 / 1M
Output Price
$0.600 / 1M
Cache Read Price
$0.003 / 1M
Cache Write Price
N/A
Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.
Flaky tests
5
Flaky tests had mixed outcomes across runs (at least one pass and one fail).
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
#100 DeepSeek V4.1 Flash 1017
max- Cost
- $0.018
- Time
- 135.0s
- Tokens
- 35,163 tok
Charts
Choose the first model, then click a second model to open a side-by-side page.
Top Models by Score
Score vs Total Cost
Response Time (avg)
Score vs Response Time (avg)
Total Output Tokens
Score vs Total Output Tokens
Quick Compare
DeepSeek V4.1 Flash 1017maxvsGemini 4 ArgonlowDeepSeek V4.1 Flash 1017maxvsGrok 4.5highDeepSeek V4.1 Flash 1017maxvsQwen3.7 PlusmediumDeepSeek V4.1 Flash 1017maxvsGLM 5.2mediumDeepSeek V4.1 Flash 1017maxvsLing 3.1 FlashmediumDeepSeek V4.1 Flash 1017maxvsGPT-5.6 TerramediumDeepSeek V4.1 Flash 1017maxvsClaude Opus 4.8mediumDeepSeek V4.1 Flash 1017maxvsSeed-2.0-LitemediumDeepSeek V4.1 Flash 1017maxvsGPT-6 LunamediumDeepSeek V4.1 Flash 1017maxvsSeed-2.0-Codelow
Category Breakdown
| Category | Score | Consistency | Tests Correct |
|---|---|---|---|
| Agentic | 10.0 | 10.0 | |
| Anti-AI Tricks | 8.3 | 10.0 | |
| Coding | 5.6 | 4.2 | |
| Combined | 10.0 | 10.0 | |
| Data parsing and extraction | 10.0 | 10.0 | |
| Domain specific | 5.9 | 7.2 | |
| General Intelligence | 10.0 | 10.0 | |
| Instructions following | 10.0 | 10.0 | |
| Puzzle Solving | 6.3 | 4.9 | |
| Tool Calling | 10.0 | 10.0 | |
| Trivia | 3.0 | 10.0 |