ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Hy3 preview vs DiffusionGemma 26B A4B IT vs Mistral Small 4

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

Price in rating: Included · Exclude price

  1. Tencent

    Hy3 preview

    Released Apr 20, 2026

    55/100
    • ECI—
    • Price$0.18 / $0.60
    • Context256K
  2. Google

    DiffusionGemma 26B A4B IT

    Released Jun 9, 2026

    76/100
    • ECI—
    • Price$0.05 / $0.15
    • Context262K
  3. Mistral AI

    Mistral Small 4

    Released Mar 16, 2026

    67/100
    • ECI—
    • Price$0.15 / $0.60
    • Context256K
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (DiffusionGemma 26B A4B IT 76/100, Mistral Small 4 67/100, Hy3 preview 55/100) compare the available specifications and price. The score weighs price 50%, inputs & features 30%, context window 20%. 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 priceDiffusionGemma 26B A4B ITDiffusionGemma 26B A4B IT $0.075 · Mistral Small 4 $0.263 · Hy3 preview $0.285 per 1M tokens (3:1 blend)
  • Longest contextDiffusionGemma 26B A4B ITDiffusionGemma 26B A4B IT 262,144 · Hy3 preview 256,000 · Mistral Small 4 256,000 tokens
  • Widest inputsDiffusionGemma 26B A4B IT and Mistral Small 4Hy3 preview: Text · DiffusionGemma 26B A4B IT: Text, Images · Mistral Small 4: Text, Images
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightHy3 previewDiffusionGemma 26B A4B ITMistral Small 4
Price50%729974
Inputs & features30%356070
Context window20%434343
Specs & price only100%55/10076/10067/100

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

02 — Side by side

Every spec in one table

Highlighted cells lead their row. Dashes mean the data is not published.

Hy3 preview vs DiffusionGemma 26B A4B IT vs Mistral Small 4 specifications side by side
SpecificationHy3 previewTencentDiffusionGemma 26B A4B ITGoogleMistral Small 4Mistral AI
Capability
Capabilities Index (ECI)———
ECI rank———
Price per million tokens
Input$0.18$0.05 (best)$0.15
Output$0.60$0.15 (best)$0.60
Cached input——$0.015
Blended (3:1)$0.285$0.075 (best)$0.263
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 3 providersMedian of 1 providersOfficial Mistral API
Limits
Context window256,000 tokens262,144 tokens (best)256,000 tokens
Max output64,000 tokens32,768 tokens256,000 tokens (best)
Inputs and features
TextYesYesYes
ImagesNoYesYes
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningYeslow · medium · highYesYeshigh
Tool callingYesYesYes
Structured outputNoNoYes
Availability
WeightsOpenOpenApache 2.0Open
API model IDhy3-preview—mistral-small-2603
API providers5320 (best)
ReleasedApr 20, 2026Jun 9, 2026Mar 16, 2026
Knowledge cutoff—Jan 2025Jun 2025
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.

  • Hy3 preview$3.00
  • DiffusionGemma 26B A4B IT$0.80
  • Mistral Small 4$2.70
04 — Questions

Which should you choose?

Which is better: Hy3 preview, DiffusionGemma 26B A4B IT or Mistral Small 4?

There is not enough shared capability evidence to name an overall winner. The scores (DiffusionGemma 26B A4B IT 76/100, Mistral Small 4 67/100, Hy3 preview 55/100) compare the available specifications and price. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.

Which is cheaper, Hy3 preview, DiffusionGemma 26B A4B IT or Mistral Small 4?

DiffusionGemma 26B A4B IT is cheaper at $0.05 input / $0.15 output per million tokens (median across 1 API provider). Mistral Small 4 costs $0.15 input / $0.60 output per million tokens (official Mistral API price); Hy3 preview costs $0.18 input / $0.60 output per million tokens (median across 3 API providers; free on Tencent TokenHub). At a typical mix of three input tokens to one output token, that is $0.075 per million tokens for DiffusionGemma 26B A4B IT versus $0.263 for Mistral Small 4 (3.5× as much) and $0.285 for Hy3 preview (3.8× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. Hy3 preview has not been scored yet, DiffusionGemma 26B A4B IT has not been scored yet and Mistral Small 4 has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for Hy3 preview, DiffusionGemma 26B A4B IT and Mistral Small 4 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?

DiffusionGemma 26B A4B IT has the largest context window at 262,144 tokens, against 256,000 for Hy3 preview and 256,000 for Mistral Small 4. Maximum output per response: Hy3 preview up to 64,000, DiffusionGemma 26B A4B IT up to 32,768, Mistral Small 4 up to 256,000 tokens.

Which can read images, PDFs, audio or video?

Hy3 preview accepts text; DiffusionGemma 26B A4B IT accepts text and images; Mistral Small 4 accepts text and images. DiffusionGemma 26B A4B IT handles the widest range of inputs.

Are any of these open source?

Yes, all three publish their weights (Apache 2.0), so you can self-host them.

Which is newer?

DiffusionGemma 26B A4B IT is the newest, released Jun 9, 2026. Hy3 preview came out Apr 20, 2026; Mistral Small 4 came out Mar 16, 2026. Knowledge cutoff: DiffusionGemma 26B A4B IT Jan 2025, Mistral Small 4 Jun 2025.

How do you decide the winner?

This rating uses price (50%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (30%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (20%, 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 18, 2026.