Sign Up

Decision support

DeepSeek V4.1 Flash 1017 vs MiMo-V2.6-Flash

Compare canonical objects directly. The useful dimensions change with the object type, so model and hardware comparisons stay separate.

ModelsHardware
DeepSeek V4.1 Flash 1017

DeepSeek

Sources:models.devOpenRouter

MiMo-V2.6-Flash

Xiaomi

Sources:models.devOpenRouter

Benchmark results

Head to head on 2 benchmarks both models report

  • DeepSeek V4.1 Flash 10172leads
  • MiMo-V2.6-Flash0leads

  • Arena Text

    Arena rating · Higher is better

    • DeepSeek V4.1 Flash 1017
      Not reported
    • MiMo-V2.6-Flash
      1448
      Arena #86CI 1443 to 1453
    Runs and sources
    DeepSeek V4.1 Flash 1017

    Not reported

    MiMo-V2.6-Flash
    • default144895% CI 1443 to 1453 · 22,354 votes
    Source:Arena Text leaderboard, 2026-11-09

  • Arena WebDev

    Arena rating · Higher is better

    • DeepSeek V4.1 Flash 1017
      Best1664
      +26 ahead · Arena #17CI 1657 to 1670
    • MiMo-V2.6-Flash
      1638
      Arena #24CI 1630 to 1645
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • max166495% CI 1657 to 1670 · 13,952 votes
    MiMo-V2.6-Flash
    • default163895% CI 1630 to 1645 · 8,890 votes
    Source:Arena WebDev leaderboard, 2026-11-10

  • Arena Vision

    Arena rating · Higher is better

    • DeepSeek V4.1 Flash 1017
      Not reported
    • MiMo-V2.6-Flash
      1248
      Arena #56CI 1239 to 1256
    Runs and sources
    DeepSeek V4.1 Flash 1017

    Not reported

    MiMo-V2.6-Flash
    • default124895% CI 1239 to 1256 · 6,906 votes
    Source:Arena Vision leaderboard, 2026-11-09

  • Agent Arena

    IPS score · Higher is better

    • DeepSeek V4.1 Flash 1017
      Best+0.020
      +0.020 ahead · Arena #22CI +0.015 to +0.025
    • MiMo-V2.6-Flash
      -0.000
      Arena #33CI -0.011 to +0.010
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • max+0.02095% CI +0.015 to +0.025 · 24,468,135 observations
    MiMo-V2.6-Flash
    • default-0.00095% CI -0.011 to +0.010 · 3,948,227 observations
    Source:Agent Arena leaderboard, 2026-11-11

  • ARC-AGI-2

    Best published score · Higher is better

    • DeepSeek V4.1 Flash 1017
      78.1%
      #20 of 85
    • MiMo-V2.6-Flash
      Not reported
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • none2.1%
    • low73.6%
    • high68.6%
    • max78.1%
    MiMo-V2.6-Flash

    Not reported

    Source:External evaluation via Epoch AIOriginal-source terms apply

  • FrontierMath-Tier-4-v2-Private

    Best published score · Higher is better

    • DeepSeek V4.1 Flash 1017
      46.3%
      #23 of 72
    • MiMo-V2.6-Flash
      Not reported
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • max46.3%
    MiMo-V2.6-Flash

    Not reported

    Source:Epoch AI evaluation

  • FrontierMath-Tiers-1-3-v2-Private

    Best published score · Higher is better

    • DeepSeek V4.1 Flash 1017
      71.2%
      #28 of 87
    • MiMo-V2.6-Flash
      Not reported
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • max71.2%
    MiMo-V2.6-Flash

    Not reported

    Source:Epoch AI evaluation

  • GPQA diamond

    Best published score · Higher is better

    • DeepSeek V4.1 Flash 1017
      93.2%
      #20 of 161
    • MiMo-V2.6-Flash
      Not reported
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • max93.2%
    MiMo-V2.6-Flash

    Not reported

    Source:Epoch AI evaluation

  • OTIS Mock AIME 2024-2025

    Best published score · Higher is better

    • DeepSeek V4.1 Flash 1017
      99.4%
      #15 of 158
    • MiMo-V2.6-Flash
      Not reported
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • max99.4%
    MiMo-V2.6-Flash

    Not reported

    Source:Epoch AI evaluation

  • SimpleQA Verified

    Best published score · Higher is better

    • DeepSeek V4.1 Flash 1017
      42.5%
      #48 of 87
    • MiMo-V2.6-Flash
      Not reported
    Runs and sources
    DeepSeek V4.1 Flash 1017
    • max42.5%
    MiMo-V2.6-Flash

    Not reported

    Source:Epoch AI evaluation

Each model's best published result is shown; "Not reported" means no matched evaluation, not a score of zero. Arena and DeepSWE ranks are the leaderboards' own; other ranks are among catalogue models with a score. Results from different harnesses or settings are not directly comparable.

Original-source terms apply to evaluations credited to their original publishers.

Benchmark compilation: Epoch AI, Capabilities & benchmarking (CC BY attribution). Scores are converted to percentages and model names matched to the Vertical Tabs catalog. Individual evaluations are credited to their original sources. External results may carry separate licences and evaluation configurations; do not interpret scores from different harnesses as directly comparable.

Arena ratings: Arena (lmarena-ai/leaderboard-dataset), licensed CC BY 4.0. Ratings come from pairwise human votes and are only comparable within one leaderboard; the interval is Arena's 95% confidence interval. Model names are matched to the Vertical Tabs catalog.

Model specifications

Release date
October 17, 2026
September 22, 2026
Context window
1M
1.1M
Native context
—
1M
Max output tokens
393,216
131,072
Knowledge cutoff
2025-05
—
Open weights
Open weights
Open weights
Parameter count
763B
161B

API pricing · USD per 1M tokens

Input
$0.15
$0.14
Output
$0.60
$0.28
Cached input
$0.003
$0.0028
Cache write
—
—

Capabilities and modalities

Input modalities
text, image
text, image, video, audio
Output modalities
text
text
Tool calling
Supported
Supported
Reasoning option
Supported
Supported
Supported options
frequency_penalty, include_reasoning, logit_bias, logprobs, max_completion_tokens, max_tokens, min_p, parallel_tool_calls, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p
frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, parallel_tool_calls, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p
Capabilities
—
completion, thinking, tools, vision

Architecture and availability

Architecture
—
mimo_v2
Model families
—
mimo_v2
Tokenizer
DeepSeek
Other
Instruction format
—
—
License
MIT
—
Providers
38
8
Provider endpoints
96
8
Ollama variants
0
1
Metadata source
ai-research
models.dev

A dash means the field is not reported. Provider-specific performance and pricing may differ.

To give you the best experience, we use cookies for personalization, insights, and performance. Learn more.