Qwen3.8 Flash-Next NVFP4
- Input
- —
- Output
- —
- Speed
- ~180tok/sest.
- Context
- 262Ktokens
Rankings
Coding
Lab-reported#121 / 137rank by quality23.5±10.2quality / 1001/7 benchmarks measuredWriting & chat
Lab-reported#24 / 162rank by quality74.1±9.7quality / 1001/5 benchmarks measuredVision
Lab-reported#23 / 98rank by quality71.9±8.7quality / 1001/6 benchmarks measuredHard reasoning
Lab-reported#46 / 218rank by quality68.6±6.6quality / 1002/4 benchmarks measuredAgents
Lab-reported#36 / 187rank by quality64.9±10.1quality / 1001/7 benchmarks measured
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
SWE-bench VerifiedEstimated
~#20 of 21 (estimated)~55.2%
- #1Claude Opus 596.0%
- #2Claude Sonnet 585.2%
- #19NVIDIA Nemotron-Labs Dory1.0 4B59.5%
- ~#20Qwen3.8 Flash-Next NVFP4Estimated55.2%
- #20LFM2.5-24B-A2B51.0%
No published score — estimated from related benchmarks
SWE-bench Pro (Public)Estimated
~#23 of 24 (estimated)~48.1%
- #1Claude Fable 5.293.0%
- #2Claude Opus 5.589.9%
- #22Muse Glimmer51.2%
- ~#23Qwen3.8 Flash-Next NVFP4Estimated48.1%
- #23Nex-N2.5-Mini43.8%
No published score — estimated from related benchmarks
DeepSWE v1.1Estimated
~#45 of 46 (estimated)~23.0%
- #1Gemini 4 Argon78.8%
- #2DeepSeek V4.1 Flash 101777.0%
- #44Nemotron 3.5 Super VL24.6%
- ~#45Qwen3.8 Flash-Next NVFP4Estimated23.0%
- #45Nemotron 3.5 Super VL NVFP418.2%
No published score — estimated from related benchmarks
Terminal-Bench 4.0Estimated
~#36 of 102 (estimated)~12.8%
- #1Kimi K3.1 Pro67.2%
- #2Claude Sonnet 5.563.6%
- #34DeepSeek V4 Pro (0813)14.1%
- ~#36Qwen3.8 Flash-Next NVFP4Estimated12.8%
- #36GPT-5.5 Instant (June 2026)12.6%
No published score — estimated from related benchmarks
LiveCodeBenchEstimated
~#19 of 28 (estimated)~52.2%
- #1Qwen3.8-Omni-Flash92.6%
- #2Qwen3.8 Flash-Next91.9%
- #18Phi-4-reasoning-plus53.1%
- ~#19Qwen3.8 Flash-Next NVFP4Estimated52.2%
- #19Falcon-H1-34B-Instruct49.7%
No published score — estimated from related benchmarks
SciCodeLab-reported
#115 of 11818.8%
- #1Claude Fable 5.267.7%
- #2Claude Opus 5.566.9%
- #114Ministral 3 8B20.7%
- #115Qwen3.8 Flash-Next NVFP418.8%
- #116Ministral 3 3B15.3%
huggingface.co · 2026-09-29
AA Coding Agent IndexEstimated
~#23 of 26 (estimated)~35.4
- #1Claude Opus 5.561.7
- #2Claude Opus 559.7
- #22Claude Haiku 5.537.3
- ~#23Qwen3.8 Flash-Next NVFP4Estimated35.4
- #23Gemini 3.8 Flash33.3
No published score — estimated from related benchmarks
Writing & chat
IFBenchLab-reported
#6 of 13981.0%
- #1MiniMax-M382.9%
- #2Qwen3.8 2.4T A95B82.8%
- #5Qwen3.8 Flash-Next81.3%
- #6Qwen3.8 Flash-Next NVFP481.0%
- #7Intern-S2-Preview (35B-A3B)80.9%
huggingface.co · 2026-09-29
LMArena Text, Non-English (Elo)Estimated
~#21 of 57 (estimated)~1445 Elo
- #1Gemini 4 ArgonFew votes1512
- #2Claude Opus 5.51497
- #19Grok 4.61447
- ~#21Qwen3.8 Flash-Next NVFP4Estimated1445
- #21Gemini 3.5 Flash-Lite1445
No published score — estimated from related benchmarks
LMArena Text (Elo)Estimated
~#20 of 57 (estimated)~1457 Elo
- #1Gemini 4 Argon1525
- #2Claude Opus 5.51507
- #19Claude Sonnet 51461
- ~#20Qwen3.8 Flash-Next NVFP4Estimated1457
- #20GPT-6 Sol1457
No published score — estimated from related benchmarks
EQ-Bench Creative Writing v3 (Elo)Estimated
~#9 of 26 (estimated)~1921 Elo
- #1GPT-6 Astra2173
- #2Claude Fable 5.12162
- #8Grok 4.72007
- ~#9Qwen3.8 Flash-Next NVFP4Estimated1921
- #9Muse Spark 1.31906
No published score — estimated from related benchmarks
Vision
MMMU-ProLab-reported
#24 of 8778.3%
- #1Claude Opus 5.587.7%
- #2GPT-6 Astra86.9%
- #23GPT-5.3 Codex78.5%
- #24Qwen3.8 Flash-Next NVFP478.3%
- #25Grok 4.778.2%
huggingface.co · 2026-09-29
CharXiv ReasoningEstimated
~#6 of 12 (estimated)~82.4%
- #1Gemini 3.8 Flash86.2%
- #2Kimi K384.8%
- #5MiniMax-M3.1-Flash82.5%
- ~#6Qwen3.8 Flash-Next NVFP4Estimated82.4%
- #6Nemotron 3.5 Super VL80.8%
No published score — estimated from related benchmarks
OmniDocBench v1.5Estimated
~#4 of 8 (estimated)~88.7%
- #1MiniMax-M391.6%
- #2Kimi K391.1%
- #2Qwen3.8 27B91.1%
- ~#4Qwen3.8 Flash-Next NVFP4Estimated88.7%
- #4Gemini 3.1 Pro85.3%
No published score — estimated from related benchmarks
GDP.pdf (AA)Estimated
~#20 of 48 (estimated)~18.6%
- #1Muse Wonder Full32.6%
- #2GPT-6 Astra31.0%
- #19MiMo-V2.6-Pro19.2%
- ~#20Qwen3.8 Flash-Next NVFP4Estimated18.6%
- #20Gemini 3.1 Pro17.8%
No published score — estimated from related benchmarks
ChartographyEstimated
~#9 of 25 (estimated)~31.4%
- #1Claude Fable 5.273.4%
- #2GPT-6 Astra71.0%
- #8GPT-5.6 Terra34.0%
- ~#9Qwen3.8 Flash-Next NVFP4Estimated31.4%
- #9GPT-5.531.0%
No published score — estimated from related benchmarks
Hard reasoning
GPQA DiamondLab-reported
#22 of 19291.5%
- #1GPT-6 Astra96.1%
- #2Gemini 3.8 Flash95.3%
- #22GPT-5.3 Codex91.5%
- #22Qwen3.8 Flash-Next NVFP491.5%
- #24DeepSeek V4.1 Flash NVFP491.3%
huggingface.co · 2026-09-29
Humanity's Last ExamLab-reported
#45 of 19835.4%
- #1Claude Fable 5.274.3%
- #2Claude Opus 5.561.4%
- #44Qwen3.7 Plus35.6%
- #45Qwen3.8 Flash-Next NVFP435.4%
- #46Kimi K2.7 Code35.0%
huggingface.co · 2026-09-29
AIME (latest)Estimated
~#2 of 34 (estimated)~99.1%
No published score — estimated from related benchmarks
AA Intelligence IndexEstimated
~#37 of 103 (estimated)~34.6
- #1Gemini 4 Carbon59.4
- #2Claude Opus 5.557.6
- #36DeepSeek V4 Pro (0813)36.0
- ~#37Qwen3.8 Flash-Next NVFP4Estimated34.6
- #37Qwen3.8 27B33.7
No published score — estimated from related benchmarks
Agents
τ²-bench (Telecom)Lab-reported
#13 of 11990.1%
- #1Step 3.7 Flash98.5%
- #2DeepSeek-V4-Pro-0813-nvfp4-DSpark98.3%
- #13Kimi K2.7 Code90.1%
- #13Qwen3.8 Flash-Next NVFP490.1%
- #13Trinity Large Thinking90.1%
huggingface.co · 2026-09-29
OSWorld-VerifiedEstimated
~#9 of 15 (estimated)~74.5%
- #1Kimi K384.8%
- #2Qwen3.8 27B84.3%
- #8GPT-5.578.7%
- ~#9Qwen3.8 Flash-Next NVFP4Estimated74.5%
- #9Gemini 3.5 Flash-Lite74.0%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#3 of 5 (estimated)~53.8%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#14 of 22 (estimated)~79.4%
- #1Atria Dawn Preview92.5%
- #2GPT-6 Astra91.5%
- #13Nex-N2.5-Mini83.4%
- ~#14Qwen3.8 Flash-Next NVFP4Estimated79.4%
- #14Nemotron 3.5 Super VL78.0%
No published score — estimated from related benchmarks
GDPval-AA (Elo)Estimated
~#21 of 92 (estimated)~1530 Elo
- #1Claude Opus 5.51866
- #2Muse Wonder Full1857
- #20Kimi K31533
- ~#21Qwen3.8 Flash-Next NVFP4Estimated1530
- #21GPT-6 Sol1487
No published score — estimated from related benchmarks
AutomationBench-AAEstimated
~#29 of 85 (estimated)~39.0%
- #1Claude Fable 5.276.4%
- #2Muse Wonder Full71.3%
- #28GPT-5.547.3%
- ~#29Qwen3.8 Flash-Next NVFP4Estimated39.0%
- #29K2 Horizon 375B A23B37.2%
No published score — estimated from related benchmarks
τ-Bench Banking (AA)Estimated
~#22 of 91 (estimated)~29.2%
- #1Muse Spark 1.350.5%
- #2GLM-5.350.3%
- #21K2 Horizon MoVA 36B A4B32.0%
- ~#22Qwen3.8 Flash-Next NVFP4Estimated29.2%
- #22Inkling29.1%
No published score — estimated from related benchmarks