ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Granite-4.0-H-Small vs Ministral 3 8B vs olmOCR 2 7B

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. IBM

    Granite-4.0-H-Small

    Released Oct 2, 2025

    62/100
    • ECI—
    • Price$0.064 / $0.265
    • Context131K
  2. Mistral AI

    Ministral 3 8B

    Released Dec 2, 2025

    69/100
    • ECI—
    • Price$0.15 / $0.15
    • Context262K
  3. Allenai

    olmOCR 2 7B

    Released Oct 22, 2025

    51/100
    • ECI—
    • Price$0.071 / $0.345
    • Context16K
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (Ministral 3 8B 69/100, Granite-4.0-H-Small 62/100, olmOCR 2 7B 51/100) 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 priceGranite-4.0-H-SmallGranite-4.0-H-Small $0.114 · olmOCR 2 7B $0.14 · Ministral 3 8B $0.15 per 1M tokens (3:1 blend)
  • Longest contextMinistral 3 8BMinistral 3 8B 262,144 · Granite-4.0-H-Small 131,072 · olmOCR 2 7B 16,384 tokens
  • Widest inputsMinistral 3 8B and olmOCR 2 7BGranite-4.0-H-Small: Text · Ministral 3 8B: Text, Images · olmOCR 2 7B: Text, Images
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightGranite-4.0-H-SmallMinistral 3 8BolmOCR 2 7B
Price50%918587
Inputs & features30%356025
Context window20%32430
Specs & price only100%62/10069/10051/100

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.

Granite-4.0-H-Small vs Ministral 3 8B vs olmOCR 2 7B specifications side by side
SpecificationGranite-4.0-H-SmallIBMMinistral 3 8BMistral AIolmOCR 2 7BAllenai
Capability
Capabilities Index (ECI)———
ECI rank———
Price per million tokens
Input$0.064 (best)$0.15$0.071
Output$0.265$0.15 (best)$0.345
Cached input—$0.015—
Blended (3:1)$0.114 (best)$0.15$0.14
Long-context rateSame rateSame rateSame rate
Price sourceOfficial watsonx.ai APIOfficial Mistral APIMedian of 1 providers
Limits
Context window131,072 tokens262,144 tokens (best)16,384 tokens
Max output131,072 tokens262,144 tokens (best)—
Inputs and features
TextYesYesYes
ImagesNoYesYes
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningNoNoNo
Tool callingYesYesNo
Structured outputYesYesNo
Availability
WeightsOpenOpenApache 2.0OpenApache-2.0
API model IDibm/granite-4-h-smallministral-8b-2512—
API providers112 (best)1
ReleasedOct 2, 2025Dec 2, 2025Oct 22, 2025
Knowledge cutoff———
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.

  • Granite-4.0-H-Small$1.17
  • Ministral 3 8B$1.80
  • olmOCR 2 7B$1.40
04 — Questions

Which should you choose?

Which is better: Granite-4.0-H-Small, Ministral 3 8B or olmOCR 2 7B?

There is not enough shared capability evidence to name an overall winner. The scores (Ministral 3 8B 69/100, Granite-4.0-H-Small 62/100, olmOCR 2 7B 51/100) 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, Granite-4.0-H-Small, Ministral 3 8B or olmOCR 2 7B?

Granite-4.0-H-Small is cheaper at $0.064 input / $0.265 output per million tokens (official watsonx.ai API price). olmOCR 2 7B costs $0.071 input / $0.345 output per million tokens (median across 1 API provider); Ministral 3 8B costs $0.15 input / $0.15 output per million tokens (official Mistral API price). At a typical mix of three input tokens to one output token, that is $0.114 per million tokens for Granite-4.0-H-Small versus $0.14 for olmOCR 2 7B (1.2× as much) and $0.15 for Ministral 3 8B (1.3× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. Granite-4.0-H-Small has not been scored yet, Ministral 3 8B has not been scored yet and olmOCR 2 7B has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for Granite-4.0-H-Small, Ministral 3 8B and olmOCR 2 7B yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. Note that olmOCR 2 7B does not support tool calling, which most coding agents need.

Which has the bigger context window?

Ministral 3 8B has the largest context window at 262,144 tokens, against 131,072 for Granite-4.0-H-Small and 16,384 for olmOCR 2 7B. Maximum output per response: Granite-4.0-H-Small up to 131,072, Ministral 3 8B up to 262,144 tokens.

Which can read images, PDFs, audio or video?

Granite-4.0-H-Small accepts text; Ministral 3 8B accepts text and images; olmOCR 2 7B accepts text and images. Ministral 3 8B handles the widest range of inputs.

Are any of these open source?

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

Which is newer?

Ministral 3 8B is the newest, released Dec 2, 2025. olmOCR 2 7B came out Oct 22, 2025; Granite-4.0-H-Small came out Oct 2, 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.