ZMIME
Comparison · 3 models · Updated Nov 22, 2026

Nemotron 3.5 Super VL 120B A12B vs Qwen3.8 Flash Next vs Ornith 1.5 122B

No independent capability benchmark covers every model here yet. The score uses price and context window.No independent capability benchmark covers every model here yet. The score uses context window.

  1. NVIDIA

    Nemotron 3.5 Super VL 120B A12B

    Released Oct 15, 2026

    71/100
    • ECI—
    • Price$0.20 / $0.50
    • Context1.05M
  2. Alibaba (Qwen)

    Qwen3.8 Flash Next

    Released Aug 27, 2026

    65/100
    • ECI—
    • Price$0.20 / $0.50
    • Context262K
  3. DeepReinforce

    Ornith 1.5 122B

    Released Oct 20, 2026

    64/100
    • ECI—
    • Price$0.16 / $0.64
    • Context262K
  4. NVIDIA

    Nemotron 3.5 Super VL 120B A12B

    Released Oct 15, 2026

    65/100
    • ECI—
    • Price$0.20 / $0.50
    • Context1.05M
  5. Alibaba (Qwen)

    Qwen3.8 Flash Next

    Released Aug 27, 2026

    43/100
    • ECI—
    • Price$0.20 / $0.50
    • Context262K
  6. DeepReinforce

    Ornith 1.5 122B

    Released Oct 20, 2026

    43/100
    • ECI—
    • Price$0.16 / $0.64
    • Context262K
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (Nemotron 3.5 Super VL 120B A12B 71/100, Qwen3.8 Flash Next 65/100, Ornith 1.5 122B 64/100) compare the available specifications and price. The score weighs price 71%, context window 29%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.There is not enough shared capability evidence to name an overall winner. The scores (Nemotron 3.5 Super VL 120B A12B 65/100, Qwen3.8 Flash Next 43/100, Ornith 1.5 122B 43/100) compare the available specifications only. The score weighs context window 100%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.

  • CapabilityNot enough dataNo independent benchmark covers every model here yet
  • Lowest priceNemotron 3.5 Super VL 120B A12B and Qwen3.8 Flash NextNemotron 3.5 Super VL 120B A12B $0.275 · Qwen3.8 Flash Next $0.275 · Ornith 1.5 122B $0.28 per 1M tokens (3:1 blend)
  • Longest contextNemotron 3.5 Super VL 120B A12BNemotron 3.5 Super VL 120B A12B 1,048,576 · Qwen3.8 Flash Next 262,144 · Ornith 1.5 122B 262,144 tokens
  • Widest inputsNemotron 3.5 Super VL 120B A12B and Qwen3.8 Flash NextNemotron 3.5 Super VL 120B A12B: Text, Images, Video · Qwen3.8 Flash Next: Text, Images, Video · Ornith 1.5 122B: Text, Images
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightNemotron 3.5 Super VL 120B A12BQwen3.8 Flash NextOrnith 1.5 122B
Price71%737373
Context window29%654343
Specs & price only100%71/10065/10064/100

Left out because at least one model lacks the data: capability and inputs & features. The remaining weights were rescaled.

How the score is built
MeasureWeightNemotron 3.5 Super VL 120B A12BQwen3.8 Flash NextOrnith 1.5 122B
Context window100%654343
Specs only100%65/10043/10043/100

Left out because at least one model lacks the data: capability and inputs & features. The remaining weights were rescaled.

02 — Side by side

Every spec in one table

Highlighted cells lead their row. Unavailable specifications are marked “Not published”. Publisher results include their evaluation settings and source.

Nemotron 3.5 Super VL 120B A12B vs Qwen3.8 Flash Next vs Ornith 1.5 122B specifications side by side
SpecificationNemotron 3.5 Super VL 120B A12BNVIDIAQwen3.8 Flash NextAlibaba (Qwen)Ornith 1.5 122BDeepReinforce
Capability
Capabilities Index (ECI)Not publishedNot publishedNot published
ECI rankNot publishedNot publishedNot published
ECI confidence intervalNot publishedNot publishedNot published
ECI evaluation dateNot publishedNot publishedNot published
Documented capabilitiesNVIDIA vision-language reasoning model for text, image, and video understanding and agentic tool useOpen-weight experimental preview of the Qwen4 architecture: hybrid-attention MoE (125B total, 6B active) with vision encoder for coding, agent tasks, and image and video understandingMixture-of-experts coding-reasoning model for agentic software tasks, tool use, and image understanding
Model parametersNot publishedNot publishedNot published
Output modalitiesTextTextText
Publisher-reported benchmarks
DeepSWE 1.1Publisher-reported resultNot published58.7scoreReported resultNot published
SWE-Bench Pro · Qwen refined and corrected task setPublisher-reported resultNot published62.5scoreReported resultNot published
SWE-Bench MultilingualPublisher-reported resultNot published81scoreReported resultNot published
NL2RepoPublisher-reported resultNot published48.1scoreReported resultNot published
CoWorkBench · internalPublisher-reported resultNot published73.9scoreReported resultNot published
JobBenchPublisher-reported resultNot published55.7scoreReported resultNot published
Agents' Last ExamPublisher-reported resultNot published24.3pass@1Reported resultNot published
Agents' Last ExamPublisher-reported resultNot published51.2scoreReported resultNot published
Toolathlon-VerifiedPublisher-reported resultNot published73.5pass@1Reported resultNot published
IFBenchPublisher-reported resultNot published81.3scoreReported resultNot published
GPQA DiamondPublisher-reported resultNot published91.7scoreReported resultNot published
Humanity's Last Examwithout tools; GPT-4o judgeNot published35.9scoreReported resultNot published
LiveCodeBench 6Publisher-reported resultNot published91.9scoreReported resultNot published
ClawEval-MMPublisher-reported resultNot published64.4pass@3Reported resultNot published
ClawEval-MMPublisher-reported resultNot published60.4average scoreReported resultNot published
RecreationBench · internalPublisher-reported resultNot published49.9scoreReported resultNot published
AndroidWorldPublisher-reported resultNot published84.5scoreReported resultNot published
OSWorld 2.0Publisher-reported resultNot published19.4binary completion rateReported resultNot published
OSWorld 2.0Publisher-reported resultNot published52.3partial scoreReported resultNot published
Vision2Webgpt-5.4-2026-03-05 judgeNot published64scoreReported resultNot published
ERQAPublisher-reported resultNot published72.3scoreReported resultNot published
LVBenchPublisher-reported resultNot published76.6scoreReported resultNot published
RealWorldQAPublisher-reported resultNot published88.5scoreReported resultNot published
MathVision · corrected annotationswithout code interpreterNot published90.6scoreReported resultNot published
MathVision · corrected annotationswith code interpreterNot published95.7scoreReported resultNot published
CharXiv Reasoningwithout code interpreterNot published84.6scoreReported resultNot published
CharXiv Reasoningwith code interpreterNot published90.6scoreReported resultNot published
Price per million tokens
Input$0.20$0.20$0.16 (best)
Output$0.50 (best)$0.50 (best)$0.64
Cached inputNot publishedNot publishedNot published
Blended (3:1)$0.275 (best)$0.275 (best)$0.28
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 15 providersMedian of 5 providersMedian of 1 providers
Limits
Context window1,048,576 tokens (best)262,144 tokens262,144 tokens
Max inputNot publishedNot publishedNot published
Max output8,192 tokens131,072 tokens (best)32,768 tokens
Inputs and features
TextYesYesYes
ImagesYesYesYes
PDFsNoNoNo
AudioNoNoNo
VideoYesYesNo
ReasoningYesYesYes
Tool callingYesYesYes
Structured outputNot publishedNot publishedNot published
File attachmentsYesYesYes
Temperature controlYesYesYes
Reasoning controltogglealwaysalways
Availability
WeightsOpenOpenqwen-community-1.0OpenMIT
API model IDnvidia/nemotron-3.5-super-vl-120b-a12bNot publishedNot published
API providers19 (best)61
ReleasedOct 15, 2026Aug 27, 2026Oct 20, 2026
Knowledge cutoffNot publishedNot publishedNot published
Publisher documentationPublisher documentation · Model weightsModel weightsModel weights
Source metadata updatedOct 15, 2026Aug 27, 2026Oct 20, 2026
Data sourcesmodels.dev · Epoch AImodels.dev · Epoch AImodels.dev · Epoch AI
Catalog checkedNov 22, 2026Nov 22, 2026Nov 22, 2026
03 — Cost

What would a month cost?

Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.

  • Nemotron 3.5 Super VL 120B A12B$3.00
  • Qwen3.8 Flash Next$3.00
  • Ornith 1.5 122B$2.88
04 — Questions

Which should you choose?

Which is better: Nemotron 3.5 Super VL 120B A12B, Qwen3.8 Flash Next or Ornith 1.5 122B?

There is not enough shared capability evidence to name an overall winner. The scores (Nemotron 3.5 Super VL 120B A12B 71/100, Qwen3.8 Flash Next 65/100, Ornith 1.5 122B 64/100) compare the available specifications and price. The score weighs price 71%, context window 29%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.There is not enough shared capability evidence to name an overall winner. The scores (Nemotron 3.5 Super VL 120B A12B 65/100, Qwen3.8 Flash Next 43/100, Ornith 1.5 122B 43/100) compare the available specifications only. The score weighs context window 100%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.

Which is cheaper, Nemotron 3.5 Super VL 120B A12B, Qwen3.8 Flash Next or Ornith 1.5 122B?

Nemotron 3.5 Super VL 120B A12B is cheaper at $0.20 input / $0.50 output per million tokens (median across 15 API providers; free on Nvidia). Qwen3.8 Flash Next costs $0.20 input / $0.50 output per million tokens (median across 5 API providers); Ornith 1.5 122B costs $0.16 input / $0.64 output per million tokens (median across 1 API provider). At a typical mix of three input tokens to one output token, that is $0.275 per million tokens for Nemotron 3.5 Super VL 120B A12B versus $0.275 for Qwen3.8 Flash Next (1× as much) and $0.28 for Ornith 1.5 122B (1× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. Nemotron 3.5 Super VL 120B A12B has not been scored yet, Qwen3.8 Flash Next has not been scored yet and Ornith 1.5 122B has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for Nemotron 3.5 Super VL 120B A12B, Qwen3.8 Flash Next and Ornith 1.5 122B yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. All three support tool calling for agent workflows.

Which has the bigger context window?

Nemotron 3.5 Super VL 120B A12B has the largest context window at 1,048,576 tokens, against 262,144 for Qwen3.8 Flash Next and 262,144 for Ornith 1.5 122B. Maximum output per response: Nemotron 3.5 Super VL 120B A12B up to 8,192, Qwen3.8 Flash Next up to 131,072, Ornith 1.5 122B up to 32,768 tokens.

Which can read images, PDFs, audio or video?

Nemotron 3.5 Super VL 120B A12B accepts text, images and video; Qwen3.8 Flash Next accepts text, images and video; Ornith 1.5 122B accepts text and images. Nemotron 3.5 Super VL 120B A12B handles the widest range of inputs.

Are any of these open source?

Yes, all three publish their weights (qwen-community-1.0 and MIT), so you can self-host them.

Which is newer?

Ornith 1.5 122B is the newest, released Oct 20, 2026. Nemotron 3.5 Super VL 120B A12B came out Oct 15, 2026; Qwen3.8 Flash Next came out Aug 27, 2026.

How do you decide the winner?

This rating uses price (71%, blended price per million tokens (3 input : 1 output), log scale); context window (29%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 22, 2026.This rating uses context window (100%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 22, 2026.