ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Pixtral Large (25.02) vs Vision Small vs Gemini 2.5 Pro Preview 06-05

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. Mistral AI

    Pixtral Large (25.02)

    Released Apr 8, 2025

    34/100
    • ECI—
    • Price$2.00 / $6.00
    • Context128K
  2. Vispark

    Vision Small

    Released May 15, 2024

    62/100
    • ECI—
    • Price$1.05 / $3.16
    • Context1M
  3. Google

    Gemini 2.5 Pro Preview 06-05

    Released Jun 5, 2025

    49/100
    • ECI—
    • Price$1.25 / $9.75
    • Context1.05M
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (Vision Small 62/100, Gemini 2.5 Pro Preview 06-05 49/100, Pixtral Large (25.02) 34/100) 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 priceVision SmallVision Small $1.58 · Pixtral Large (25.02) $3.00 · Gemini 2.5 Pro Preview 06-05 $3.38 per 1M tokens (3:1 blend)
  • Longest contextGemini 2.5 Pro Preview 06-05Gemini 2.5 Pro Preview 06-05 1,048,576 · Vision Small 1,000,000 · Pixtral Large (25.02) 128,000 tokens
  • Widest inputsVision SmallPixtral Large (25.02): Text, Images · Vision Small: Text, Images, PDFs, Audio, Video · Gemini 2.5 Pro Preview 06-05: Text, Images, PDFs, Audio
  • Self-hostingNo confirmed open weightsAll three are available only through APIs
How the score is built
MeasureWeightPixtral Large (25.02)Vision SmallGemini 2.5 Pro Preview 06-05
Price50%263924
Inputs & features30%5010080
Context window20%326465
Specs & price only100%34/10062/10049/100

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.

Pixtral Large (25.02) vs Vision Small vs Gemini 2.5 Pro Preview 06-05 specifications side by side
SpecificationPixtral Large (25.02)Mistral AIVision SmallVisparkGemini 2.5 Pro Preview 06-05Google
Capability
Capabilities Index (ECI)———
ECI rank———
Price per million tokens
Input$2.00$1.05 (best)$1.25
Output$6.00$3.16 (best)$9.75
Cached input———
Blended (3:1)$3.00$1.58 (best)$3.38
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 3 providersOfficial Vispark APIMedian of 4 providers
Limits
Context window128,000 tokens1,000,000 tokens1,048,576 tokens (best)
Max output8,192 tokens65,536 tokens (best)65,536 tokens (best)
Inputs and features
TextYesYesYes
ImagesYesYesYes
PDFsNoYesYes
AudioNoYesYes
VideoNoYesNo
ReasoningNoYesYes
Tool callingYesYesYes
Structured outputNoYesNo
Availability
WeightsProprietaryProprietaryProprietary
API model ID—vispark/vision-small—
API providers314 (best)
ReleasedApr 8, 2025May 15, 2024Jun 5, 2025
Knowledge cutoff——Jan 31, 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.

  • Pixtral Large (25.02)$32.00
  • Vision Small$16.82
  • Gemini 2.5 Pro Preview 06-05$32.00
04 — Questions

Which should you choose?

Which is better: Pixtral Large (25.02), Vision Small or Gemini 2.5 Pro Preview 06-05?

There is not enough shared capability evidence to name an overall winner. The scores (Vision Small 62/100, Gemini 2.5 Pro Preview 06-05 49/100, Pixtral Large (25.02) 34/100) 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, Pixtral Large (25.02), Vision Small or Gemini 2.5 Pro Preview 06-05?

Vision Small is cheaper at $1.05 input / $3.16 output per million tokens (official Vispark API price). Pixtral Large (25.02) costs $2.00 input / $6.00 output per million tokens (median across 3 API providers); Gemini 2.5 Pro Preview 06-05 costs $1.25 input / $9.75 output per million tokens (median across 4 API providers). At a typical mix of three input tokens to one output token, that is $1.58 per million tokens for Vision Small versus $3.00 for Pixtral Large (25.02) (1.9× as much) and $3.38 for Gemini 2.5 Pro Preview 06-05 (2.1× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. Pixtral Large (25.02) has not been scored yet, Vision Small has not been scored yet and Gemini 2.5 Pro Preview 06-05 has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for Pixtral Large (25.02), Vision Small and Gemini 2.5 Pro Preview 06-05 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?

Gemini 2.5 Pro Preview 06-05 has the largest context window at 1,048,576 tokens, against 1,000,000 for Vision Small and 128,000 for Pixtral Large (25.02). Maximum output per response: Pixtral Large (25.02) up to 8,192, Vision Small up to 65,536, Gemini 2.5 Pro Preview 06-05 up to 65,536 tokens.

Which can read images, PDFs, audio or video?

Pixtral Large (25.02) accepts text and images; Vision Small accepts text, images, PDFs, audio and video; Gemini 2.5 Pro Preview 06-05 accepts text, images, PDFs and audio. Vision Small handles the widest range of inputs.

Are any of these open source?

No. Pixtral Large (25.02), Vision Small and Gemini 2.5 Pro Preview 06-05 are proprietary and only available through APIs and apps.

Which is newer?

Gemini 2.5 Pro Preview 06-05 is the newest, released Jun 5, 2025. Pixtral Large (25.02) came out Apr 8, 2025; Vision Small came out May 15, 2024. Knowledge cutoff: Gemini 2.5 Pro Preview 06-05 Jan 31, 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.