ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Phi-4-multimodal-instruct vs Saba vs Command R7B Arabic

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. Microsoft

    Phi-4-multimodal-instruct

    Released Feb 26, 2025

    73/100
    • ECI—
    • Price$0.07 / $0.11
    • Context131K
  2. Mistral AI

    Saba

    Released Feb 17, 2025

    48/100
    • ECI—
    • Price$0.20 / $0.60
    • Context33K
  3. Cohere

    Command R7B Arabic

    Released Feb 27, 2025

    67/100
    • ECI—
    • Price$0.037 / $0.15
    • Context128K
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (Phi-4-multimodal-instruct 73/100, Command R7B Arabic 67/100, Saba 48/100) 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 priceCommand R7B ArabicCommand R7B Arabic $0.066 · Phi-4-multimodal-instruct $0.08 · Saba $0.30 per 1M tokens (3:1 blend)
  • Longest contextPhi-4-multimodal-instructPhi-4-multimodal-instruct 131,072 · Command R7B Arabic 128,000 · Saba 32,768 tokens
  • Widest inputsPhi-4-multimodal-instructPhi-4-multimodal-instruct: Text, Images, Audio · Saba: Text, PDFs · Command R7B Arabic: Text
  • Self-hostingPhi-4-multimodal-instruct and Command R7B ArabicPublishes downloadable weights (MIT)
How the score is built
MeasureWeightPhi-4-multimodal-instructSabaCommand R7B Arabic
Price50%9871100
Inputs & features30%603535
Context window20%321132
Specs & price only100%73/10048/10067/100

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.

Phi-4-multimodal-instruct vs Saba vs Command R7B Arabic specifications side by side
SpecificationPhi-4-multimodal-instructMicrosoftSabaMistral AICommand R7B ArabicCohere
Capability
Capabilities Index (ECI)———
ECI rank———
Price per million tokens
Input$0.07$0.20$0.037 (best)
Output$0.11 (best)$0.60$0.15
Cached input———
Blended (3:1)$0.08$0.30$0.066 (best)
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 1 providersMedian of 4 providersOfficial Cohere API
Limits
Context window131,072 tokens (best)32,768 tokens128,000 tokens
Max output131,072 tokens (best)26,214 tokens4,000 tokens
Inputs and features
TextYesYesYes
ImagesYesNoNo
PDFsNoYesNo
AudioYesNoNo
VideoNoNoNo
ReasoningNoNoNo
Tool callingYesYesYes
Structured outputNoNoYes
Availability
WeightsOpenMITProprietaryOpen
API model ID——command-r7b-arabic-02-2025
API providers14 (best)1
ReleasedFeb 26, 2025Feb 17, 2025Feb 27, 2025
Knowledge cutoffJun 2024Sep 30, 2024Jun 1, 2024
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.

  • Phi-4-multimodal-instruct$0.92
  • Saba$3.20
  • Command R7B Arabic$0.675
04 — Questions

Which should you choose?

Which is better: Phi-4-multimodal-instruct, Saba or Command R7B Arabic?

There is not enough shared capability evidence to name an overall winner. The scores (Phi-4-multimodal-instruct 73/100, Command R7B Arabic 67/100, Saba 48/100) 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, Phi-4-multimodal-instruct, Saba or Command R7B Arabic?

Command R7B Arabic is cheaper at $0.037 input / $0.15 output per million tokens (official Cohere API price). Phi-4-multimodal-instruct costs $0.07 input / $0.11 output per million tokens (median across 1 API provider); Saba costs $0.20 input / $0.60 output per million tokens (median across 4 API providers). At a typical mix of three input tokens to one output token, that is $0.066 per million tokens for Command R7B Arabic versus $0.08 for Phi-4-multimodal-instruct (1.2× as much) and $0.30 for Saba (4.6× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. Phi-4-multimodal-instruct has not been scored yet, Saba has not been scored yet and Command R7B Arabic has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for Phi-4-multimodal-instruct, Saba and Command R7B Arabic 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?

Phi-4-multimodal-instruct has the largest context window at 131,072 tokens, against 128,000 for Command R7B Arabic and 32,768 for Saba. Maximum output per response: Phi-4-multimodal-instruct up to 131,072, Saba up to 26,214, Command R7B Arabic up to 4,000 tokens.

Which can read images, PDFs, audio or video?

Phi-4-multimodal-instruct accepts text, images and audio; Saba accepts text and PDFs; Command R7B Arabic accepts text. Phi-4-multimodal-instruct handles the widest range of inputs.

Are any of these open source?

Phi-4-multimodal-instruct and Command R7B Arabic publishes its weights (MIT) and can be self-hosted; Saba is proprietary.

Which is newer?

Command R7B Arabic is the newest, released Feb 27, 2025. Phi-4-multimodal-instruct came out Feb 26, 2025; Saba came out Feb 17, 2025. Knowledge cutoff: Phi-4-multimodal-instruct Jun 2024, Saba Sep 30, 2024, Command R7B Arabic Jun 1, 2024.

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.