[mle]/DaBench
Leaderboard

DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash

Hosteddeepseek30 tok/sopenrouterserved by MorphFP81M context
Run recordCompareOpen weights
Overall score
64

102 of 153 test cases fully passed

Watch the test

DeepSeek V4.1 Flash First Test | Better V4 Flash? | Coding with OpenCode, Frontend & Agentic Work

Watch on YouTube

Scores by task

Scores give partial credit. Passed counts fully correct answers only. Select a task to read its test cases.

  1. 46Agents9 of 21 passedp90 latency 13m 36s
  2. 86Structured Extraction18 of 21 passedp90 latency 4m 3s
  3. 51Coding10 of 21 passedp90 latency 7m 2s
  4. 76Grounded Answers16 of 21 passedp90 latency 4m 12s
  5. 89Instruction Following18 of 21 passedp90 latency 7m 29s
  6. 81Business Logic17 of 21 passedp90 latency 7m 43s
  7. 69Security & permissions13 of 21 passedp90 latency 7m 1s
  8. 17Diagnosis & investigation1 of 6 passedp90 latency 4m 44s

Run it yourself

The API call the benchmark made, and the settings behind this score.

Deployment
Hosted
API
OpenRouter
Serving provider
Morph
Precision
FP8
Context window
1M
Throughput
30 tok/s
Cost of a suite run
$0.39
Cost per case
$0.0026
Time per case
3m 35s
Measured
2026-10-03
Run this buildOpenRouterHosted
curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek/deepseek-v4.1-flash","provider":{"order":["morph/fp8"],"allow_fallbacks":false},"messages":[{"role":"user","content":"Hello"}]}'

The request the benchmark sent, pinned to the provider that served it.

One email when a new model lands on the board or in the showcase. Unsubscribe anytime.