ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Pixtral Large (25.02) vs Qwen Plus Character (Japanese) vs Sonar

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

Price in rating: Excluded · Include price

  1. Mistral AI

    Pixtral Large (25.02)

    Released Apr 8, 2025

    43/100
    • ECI—
    • Price$2.00 / $6.00
    • Context128K
  2. Alibaba (Qwen)

    Qwen Plus Character (Japanese)

    Released Jan 2024

    0/100
    • ECI—
    • Price$0.50 / $1.40
    • Context8K
  3. Perplexity

    Sonar

    Released Jan 1, 2024Deprecated

    13/100
    • ECI—
    • Price$1.00 / $1.00
    • Context128K
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (Pixtral Large (25.02) 43/100, Sonar 13/100, Qwen Plus Character (Japanese) 0/100) compare the available specifications only. The score weighs inputs & features 60%, context window 40%. 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 priceQwen Plus Character (Japanese)Qwen Plus Character (Japanese) $0.725 · Sonar $1.00 · Pixtral Large (25.02) $3.00 per 1M tokens (3:1 blend)
  • Longest contextPixtral Large (25.02) and SonarPixtral Large (25.02) 128,000 · Sonar 128,000 · Qwen Plus Character (Japanese) 8,192 tokens
  • Widest inputsPixtral Large (25.02)Pixtral Large (25.02): Text, Images · Qwen Plus Character (Japanese): Text · Sonar: Text
  • Self-hostingNo confirmed open weightsAll three are available only through APIs
How the score is built
MeasureWeightPixtral Large (25.02)Qwen Plus Character (Japanese)Sonar
Inputs & features60%5000
Context window40%32032
Specs only100%43/1000/10013/100

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 Qwen Plus Character (Japanese) vs Sonar specifications side by side
SpecificationPixtral Large (25.02)Mistral AIQwen Plus Character (Japanese)Alibaba (Qwen)SonarPerplexity
Capability
Capabilities Index (ECI)———
ECI rank———
Price per million tokens
Input$2.00$0.50 (best)$1.00
Output$6.00$1.40$1.00 (best)
Cached input———
Blended (3:1)$3.00$0.725 (best)$1.00
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 3 providersOfficial Alibaba APIOfficial Perplexity API
Limits
Context window128,000 tokens (best)8,192 tokens128,000 tokens (best)
Max output8,192 tokens (best)512 tokens4,096 tokens
Inputs and features
TextYesYesYes
ImagesYesNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningNoNoNo
Tool callingYesNoNo
Structured outputNoNoNo
Availability
WeightsProprietaryProprietaryProprietary
API model ID—qwen-plus-character-jasonar
API providers319 (best)
ReleasedApr 8, 2025Jan 2024Jan 1, 2024
Knowledge cutoff—Apr 2024Sep 1, 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
  • Qwen Plus Character (Japanese)$7.80
  • Sonar$12.00
04 — Questions

Which should you choose?

Which is better: Pixtral Large (25.02), Qwen Plus Character (Japanese) or Sonar?

There is not enough shared capability evidence to name an overall winner. The scores (Pixtral Large (25.02) 43/100, Sonar 13/100, Qwen Plus Character (Japanese) 0/100) compare the available specifications only. The score weighs inputs & features 60%, context window 40%. 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), Qwen Plus Character (Japanese) or Sonar?

Qwen Plus Character (Japanese) is cheaper at $0.50 input / $1.40 output per million tokens (official Alibaba API price). Sonar costs $1.00 input / $1.00 output per million tokens (official Perplexity API price); Pixtral Large (25.02) costs $2.00 input / $6.00 output per million tokens (median across 3 API providers). At a typical mix of three input tokens to one output token, that is $0.725 per million tokens for Qwen Plus Character (Japanese) versus $1.00 for Sonar (1.4× as much) and $3.00 for Pixtral Large (25.02) (4.1× 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, Qwen Plus Character (Japanese) has not been scored yet and Sonar has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for Pixtral Large (25.02), Qwen Plus Character (Japanese) and Sonar yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. Note that Qwen Plus Character (Japanese) and Sonar does not support tool calling, which most coding agents need.

Which has the bigger context window?

Pixtral Large (25.02) and Sonar have the largest context windows (128,000 and 128,000 tokens), against 8,192 for Qwen Plus Character (Japanese). Maximum output per response: Pixtral Large (25.02) up to 8,192, Qwen Plus Character (Japanese) up to 512, Sonar up to 4,096 tokens.

Which can read images, PDFs, audio or video?

Pixtral Large (25.02) accepts text and images; Qwen Plus Character (Japanese) accepts text; Sonar accepts text. Pixtral Large (25.02) handles the widest range of inputs.

Are any of these open source?

No. Pixtral Large (25.02), Qwen Plus Character (Japanese) and Sonar are proprietary and only available through APIs and apps.

Which is newer?

Pixtral Large (25.02) is the newest, released Apr 8, 2025. Qwen Plus Character (Japanese) came out Jan 2024; Sonar came out Jan 1, 2024. Knowledge cutoff: Qwen Plus Character (Japanese) Apr 2024, Sonar Sep 1, 2025.

How do you decide the winner?

This rating uses inputs & features (60%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (40%, 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.