MiMo-V2.6-Pro
- Input
- $0.435/M tok
- Output
- $0.87/M tok
- Blended
- $0.544/M tok · 3:1
- Speed
- 49tok/s
- Context
- 1Mtokens
- Cost per task
- $0.26/task · AA
Open weights (MIT). Writing: no LMArena/EQ-Bench entry or lab-reported writing benchmark as of 2026-09-24, so its Writing entry is fully estimated from its other categories (flagged “Estimated from other categories”).
Rankings
Coding
Mixed#15 / 137rank by quality76.0±6.6quality / 1003/7 benchmarks measuredWriting & chat
Verified#14 / 162rank by quality77.5±6.0quality / 1002/5 benchmarks measuredVision
Verified#19 / 98rank by quality74.2±9.4quality / 1001/6 benchmarks measuredHard reasoning
Verified#15 / 218rank by quality80.2±5.1quality / 1002/4 benchmarks measuredPrice frontierAgents
Mixed#12 / 187rank by quality82.3±6.4quality / 1003/7 benchmarks measured
Self-hostingEstimated
How we estimate →What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.
- Hardware
- 16× B200 180GB
- FP8 weights
- Throughput
- ~43,957 tok/s
- many requests batched
- Price
- $0.28–2.49 /M tok
- busy → light use
- On your own machine
- Too large
- needs data-center GPUs
Assumptions (5)
- FP8 weights (1020B params, 42B active per token) + 40% KV-cache headroom ≈ 1428 GB
- 16× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
- ~43,957 output tok/s aggregate at batch 1480 (bandwidth-bound); MoE compute scales with active params
- Blended 3:1 input:output; prefill ~344,250 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
~#2 of 21 (estimated)~89.6%
No published score — estimated from related benchmarks
SWE-bench Pro (Public)Estimated
~#6 of 24 (estimated)~73.1%
- #1Claude Fable 5.293.0%
- #2Claude Opus 5.589.9%
- #5Claude Opus 579.2%
- ~#6MiMo-V2.6-ProEstimated73.1%
- #6Intern-S2-397B68.5%
No published score — estimated from related benchmarks
DeepSWE v1.1Lab-reported
#10 of 4571.9%
- #1Gemini 4 Argon78.8%
- #2DeepSeek V4.1 Flash 101777.0%
- #9Grok 4.772.6%
- #10MiMo-V2.6-Pro71.9%
- #11Muse Spark 1.371.7%
huggingface.co · 2026-09-21
Terminal-Bench 4.0Verified
#17 of 10134.8%
- #1Kimi K3.1 Pro67.2%
- #2Claude Sonnet 5.563.6%
- #16GPT-5.6 Terra35.4%
- #17MiMo-V2.6-Pro34.8%
- #18DeepSeek V4.1 Flash 101733.8%
artificialanalysis.ai · 2026-09-24
LiveCodeBenchEstimated
~#10 of 28 (estimated)~87.3%
- #1Qwen3.8-Omni-Flash92.6%
- #2Qwen3.8 Flash-Next91.9%
- #9Nemotron 3.5 Super VL NVFP487.4%
- ~#10MiMo-V2.6-ProEstimated87.3%
- #10Ternary Bonsai 27B82.8%
No published score — estimated from related benchmarks
SciCodeVerified
#7 of 11860.9%
- #1Claude Fable 5.267.7%
- #2Claude Opus 5.566.9%
- #6Claude Sonnet 5.561.0%
- #7MiMo-V2.6-Pro60.9%
- #8Kimi K3.1 Pro60.8%
artificialanalysis.ai · 2026-09-24
AA Coding Agent IndexEstimated
~#13 of 26 (estimated)~48.1
- #1Claude Opus 5.561.7
- #2Claude Opus 559.7
- #12Mistral Large 448.3
- ~#13MiMo-V2.6-ProEstimated48.1
- #13DeepSeek V4.1 Flash47.3
No published score — estimated from related benchmarks
Writing & chat
IFBenchEstimated
~#19 of 140 (estimated)~77.0%
- #1MiniMax-M382.9%
- #2Qwen3.8 2.4T A95B82.8%
- #18Gemini 3.1 Pro77.1%
- ~#19MiMo-V2.6-ProEstimated77.0%
- #19Muse Glimmer77.0%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Verified
#15 of 561461 Elo
- #1Gemini 4 ArgonFew votes1512
- #2Claude Opus 5.51497
- #12GPT-6 AstraFew votes1463
- #15MiMo-V2.6-Pro1461
- #16DeepSeek V4.1 Flash1459
lmarena.ai · 2026-09-25
LMArena Text (Elo)Verified
#12 of 561480 Elo
- #1Gemini 4 Argon1525
- #2Claude Opus 5.51507
- #12GLM-5.31480
- #12MiMo-V2.6-Pro1480
- #14GPT-6 AstraFew votes1477
lmarena.ai · 2026-09-25
EQ-Bench Creative Writing v3 (Elo)Estimated
~#9 of 26 (estimated)~1963 Elo
- #1GPT-6 Astra2173
- #2Claude Fable 5.12162
- #8Grok 4.72007
- ~#9MiMo-V2.6-ProEstimated1963
- #9Muse Spark 1.31906
No published score — estimated from related benchmarks
Vision
MMMU-ProEstimated
~#12 of 88 (estimated)~81.3%
- #1Claude Opus 5.587.7%
- #2GPT-6 Astra86.9%
- #11Intern-S2-397B81.7%
- ~#12MiMo-V2.6-ProEstimated81.3%
- #12GPT-5.6 Terra80.7%
No published score — estimated from related benchmarks
CharXiv ReasoningEstimated
~#4 of 12 (estimated)~83.9%
- #1Gemini 3.8 Flash86.2%
- #2Kimi K384.8%
- #3Qwen3.8 Flash-Next84.6%
- ~#4MiMo-V2.6-ProEstimated83.9%
- #4Qwen3.8-Omni-Flash83.5%
No published score — estimated from related benchmarks
OmniDocBench v1.5Estimated
~#4 of 8 (estimated)~90.1%
- #1MiniMax-M391.6%
- #2Kimi K391.1%
- #2Qwen3.8 27B91.1%
- ~#4MiMo-V2.6-ProEstimated90.1%
- #4Gemini 3.1 Pro85.3%
No published score — estimated from related benchmarks
GDP.pdf (AA)Verified
#19 of 4719.2%
- #1Muse Wonder Full32.6%
- #2GPT-6 Astra31.0%
- #18Grok 4.720.0%
- #19MiMo-V2.6-Pro19.2%
- #20Gemini 3.1 Pro17.8%
artificialanalysis.ai · 2026-09-24
ChartographyEstimated
~#9 of 25 (estimated)~34.0%
- #1Claude Fable 5.273.4%
- #2GPT-6 Astra71.0%
- #8GPT-5.6 Terra34.0%
- ~#9MiMo-V2.6-ProEstimated34.0%
- #9GPT-5.531.0%
No published score — estimated from related benchmarks
Hard reasoning
GPQA DiamondEstimated
~#3 of 193 (estimated)~94.5%
No published score — estimated from related benchmarks
Humanity's Last ExamVerified
#12 of 19849.4%
- #1Claude Fable 5.274.3%
- #2Claude Opus 5.561.4%
- #11Muse Wonder50.1%
- #12MiMo-V2.6-Pro49.4%
- #13Muse Spark 1.348.7%
artificialanalysis.ai · 2026-09-24
AIME (latest)Estimated
~#2 of 34 (estimated)~98.9%
No published score — estimated from related benchmarks
Agents
τ²-bench (Telecom)Estimated
~#3 of 120 (estimated)~98.2%
- #1Step 3.7 Flash98.5%
- #2DeepSeek-V4-Pro-0813-nvfp4-DSpark98.3%
- ~#3MiMo-V2.6-ProEstimated98.2%
- #3Gemini 3.1 Pro95.6%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#3 of 5 (estimated)~56.9%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#6 of 22 (estimated)~89.7%
- #1Atria Dawn Preview92.5%
- #2GPT-6 Astra91.5%
- #5Nex-N2.5-Pro89.7%
- ~#6MiMo-V2.6-ProEstimated89.7%
- #6Step 5 Preview88.7%
No published score — estimated from related benchmarks
AutomationBench-AAVerified
#17 of 8458.6%
- #1Claude Fable 5.276.4%
- #2Muse Wonder Full71.3%
- #16Claude Fable 5.159.4%
- #17MiMo-V2.6-Pro58.6%
- #18Kimi K358.3%
artificialanalysis.ai · 2026-09-24
τ-Bench Banking (AA)Estimated
~#11 of 91 (estimated)~44.1%
- #1Muse Spark 1.350.5%
- #2GLM-5.350.3%
- #10Gemini 3.8 Flash44.9%
- ~#11MiMo-V2.6-ProEstimated44.1%
- #11Grok 4.643.3%
No published score — estimated from related benchmarks