Skip to content
Model Pareto

Qwen3.8 Flash-Next NVFP4

NVIDIAOpen weightsNVFP4quantized from Qwen3.8 Flash-Nexthuggingface.co ↗
Input
—
Output
—
Speed
~180tok/sest.
Context
262Ktokens

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs. No API sells this model, so this is its price on the chart.

Hardware
1× B200 180GB
NVFP4 weights
Throughput
~7,974 tok/s
many requests batched
Price
$0.079–0.70 /M tok
busy → light use
On your own machine
Mac Studio M5 Max 128GB (NVFP4, ~130 tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • NVFP4 weights (125B params, 6B active per token) + 25% KV-cache headroom ≈ 88 GB
  • 1× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
  • ~7,974 output tok/s aggregate at batch 228 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~177,188 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Coding

Writing & chat

Vision

MMMUEstimated
~#2 of 3 (estimated)~71.3%
  1. #1Claude Haiku 4.573.2%
  2. ~#2Qwen3.8 Flash-Next NVFP4Estimated71.3%
  3. #2Llama 4 Maverick69.4%
No published score — estimated from related benchmarks
~#4 of 8 (estimated)~88.7%
  1. #1MiniMax-M391.6%
  2. #2Kimi K391.1%
  3. #2Qwen3.8 27B91.1%
  4. ~#4Qwen3.8 Flash-Next NVFP4Estimated88.7%
  5. #4Gemini 3.1 Pro85.3%
No published score — estimated from related benchmarks
GDP.pdf (AA)Estimated
~#20 of 48 (estimated)~18.6%
  1. #1Muse Wonder Full32.6%
  2. #2GPT-6 Astra31.0%
  3. #19MiMo-V2.6-Pro19.2%
  4. ~#20Qwen3.8 Flash-Next NVFP4Estimated18.6%
  5. #20Gemini 3.1 Pro17.8%
No published score — estimated from related benchmarks
ChartographyEstimated
~#9 of 25 (estimated)~31.4%
  1. #1Claude Fable 5.273.4%
  2. #2GPT-6 Astra71.0%
  3. #8GPT-5.6 Terra34.0%
  4. ~#9Qwen3.8 Flash-Next NVFP4Estimated31.4%
  5. #9GPT-5.531.0%
No published score — estimated from related benchmarks

Hard reasoning

Agents

~#9 of 15 (estimated)~74.5%
  1. #1Kimi K384.8%
  2. #2Qwen3.8 27B84.3%
  3. #8GPT-5.578.7%
  4. ~#9Qwen3.8 Flash-Next NVFP4Estimated74.5%
  5. #9Gemini 3.5 Flash-Lite74.0%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#3 of 5 (estimated)~53.8%
  1. #1GLM-5.3 Flash59.1%
  2. #2Kimi K358.3%
  3. ~#3Qwen3.8 Flash-Next NVFP4Estimated53.8%
  4. #3GPT-5.6 Terra50.2%
No published score — estimated from related benchmarks
~#21 of 92 (estimated)~1530 Elo
  1. #1Claude Opus 5.51866
  2. #2Muse Wonder Full1857
  3. #20Kimi K31533
  4. ~#21Qwen3.8 Flash-Next NVFP4Estimated1530
  5. #21GPT-6 Sol1487
No published score — estimated from related benchmarks
~#22 of 91 (estimated)~29.2%
  1. #1Muse Spark 1.350.5%
  2. #2GLM-5.350.3%
  3. #21K2 Horizon MoVA 36B A4B32.0%
  4. ~#22Qwen3.8 Flash-Next NVFP4Estimated29.2%
  5. #22Inkling29.1%
No published score — estimated from related benchmarks