DeepSeek · Benchmarks

DeepSeek V4.1 Flash benchmarks

DeepSeek V4.1 Flash has 6 published scores in this dataset, from a DeepSeek release dated 2026-09-10. Its composite index is 79.8, which places it 15th of 22 ranked models and 10th among 15 open-weight models. Its strongest area is agentic (index 91.3) and its weakest is coding (index 72.1). Its best placing is on Terminal-Bench, where its 90.6 ranks 1st of 13 models with a published score.

Deploy hereDeploy DeepSeek-V4.1-Flash-UNCENSORED-FP8Official page

OpenLLM Buddy runs DeepSeek-V4.1-Flash-UNCENSORED-FP8, an uncensored fine-tune of DeepSeek V4.1 Flash, on 4× H200 141GB (564 GB VRAM). The scores below are the official base model's; a fine-tune can score slightly differently.

Category profile

Index 0–100: each score relative to the best published score on that test.

ReasoningCodingAgenticKnowledge
Reasoning
75.9#11 of 16 · 2 tests
Coding
72.1#12 of 13 · 2 tests
Agentic
91.31 test · not ranked
Knowledge
—no data

Model facts

Lab
DeepSeek
Type
OpenLLM Buddy catalog
Released
2026-09-10
Parameters
552B total, 16B active (MoE)
Context
1M
License
MIT
Runs on
4× H200 141GB

DeepSeek V4.1 Flash hardware requirements and cost

Even at 4-bit it needs about 381 GB (4× H200 (564 GB)), well beyond a home workstation, so most people run it on rented multi-GPU servers. Full BF16 precision needs about 1.3 TB (Multiple nodes (over 1.1 TB)).

Estimated GPU memory for DeepSeek V4.1 Flash
PrecisionWeightsWith headroomSmallest setup
BF16 (full precision)1.1 TB1.3 TBMultiple nodes (over 1.1 TB)
FP8 / 8-bit552 GB635 GB8× H100 (640 GB)
4-bit (GGUF Q4_K_M)331 GB381 GB4× H200 (564 GB)

Estimate: parameters × bytes per parameter, plus 15% for the KV cache and runtime at modest context lengths. Long contexts and large batches need more.

On OpenLLM Buddy: DeepSeek-V4.1-Flash-UNCENSORED-FP8 runs on 4× H200 141GB as an FP8 build, with an OpenAI-compatible API and no per-token fees. Prices start at $280 USD for 11 hours.

  • 11 hours: $280 USD
  • 24 hours: $611 USD
  • 1 week: $4271 USD
  • 1 month: $17082 USD
Deploy DeepSeek V4.1 Flash

Lighter or stronger open models

Open models ranked just above and below on the composite index, with their 4-bit memory needs.

DeepSeek V4.1 Flash speed and pricing

Measured figures link to their source; estimates are labelled.

API output speed
217 tok/sArtificial Analysis median · Artificial Analysis
Time to first token
0.60 sOpenRouter model page provider stats
API price per 1M tokens
$0.300 in · $1.20 outlab list price · DeepSeek API pricing docs
OpenLLM Buddy GPU rental
from $25.42 per hourflat rate, no per-token fees
Compare speed and price across all models

Score by test

Rank is among every model with a published score on that test.

GPQA Diamond90.9%Pass@1, max reasoning effort#10 of 19 · 5.2 behind GPT-6 AstraHumanity's Last Exam36.8%Pass@1, no tools, max effort (text-only subset: 39.1)#16 of 23 · 27.6 behind Claude Opus 5.5Humanity's Last Exam (with tools)63.9%Pass@1, with tools, max effort#5 of 17 · 6.1 behind Kimi K3.1 ProTerminal-Bench90.6%Terminal-Bench 2.1, Pass@1, max effort#1 of 13 · top scoreTerminal-Bench 4.031.2%Terminal-Bench 4.0, Pass@1, max effort#15 of 16 · 39.4 behind Claude Sonnet 5.5LMArena Text Elo1476LMArena text overall, 'DeepSeek-V4.1-Flash-20260910 (Max)', ±6, as of 2026-10-27#11 of 21 · 44 behind Gemini 4 Argon

Nearby in the composite ranking

Models with the closest composite index.

  1. 1Qwen3.6 27B85.8
  2. 2Qwen3.8 27B83.2
  3. 3GLM 5.382.7
  4. 4DeepSeek V4.1 Flash79.8
  5. 5Gemma 4 31B79.2
  6. 6Step 5 Preview78.7
  7. 7Nemotron 3 Ultra74.4

Head-to-head comparisons

Side-by-side pages built from the tests both models report.

Sources

ModelTestScoreSettingSource
DeepSeek V4.1 FlashGPQA Diamond90.9Pass@1, max reasoning effortDeepSeek-V4.1-Flash model card (Hugging Face)
DeepSeek V4.1 FlashHumanity's Last Exam36.8Pass@1, no tools, max effort (text-only subset: 39.1)DeepSeek-V4.1-Flash model card (Hugging Face)
DeepSeek V4.1 FlashHumanity's Last Exam (with tools)63.9Pass@1, with tools, max effortDeepSeek-V4.1-Flash model card (Hugging Face)
DeepSeek V4.1 FlashTerminal-Bench90.6Terminal-Bench 2.1, Pass@1, max effortDeepSeek-V4.1-Flash model card (Hugging Face)
DeepSeek V4.1 FlashTerminal-Bench 4.031.2Terminal-Bench 4.0, Pass@1, max effortDeepSeek-V4.1-Flash model card (Hugging Face)
DeepSeek V4.1 FlashLMArena Text Elo1476LMArena text overall, 'DeepSeek-V4.1-Flash-20260910 (Max)', ±6, as of 2026-10-27Arena (LMArena) text leaderboard

Scores are published results from model cards, lab reports, and public leaderboards, each linked to its source. Settings such as thinking mode or test version differ between sources and are shown next to each score. OpenLLM Buddy catalog models are uncensored fine-tunes; their scores are those of the official base model. Data collected 2026-10-30. How the index works