ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Nemotron 3 Super 120B A12B vs Mistral Small 4 vs Sarvam 105B

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. NVIDIA

    Nemotron 3 Super 120B A12B

    Released Mar 11, 2026

    56/100
    • ECI—
    • Price$0.135 / $0.625
    • Context262K
  2. Mistral AI

    Mistral Small 4

    Released Mar 16, 2026

    67/100
    • ECI—
    • Price$0.15 / $0.60
    • Context256K
  3. Sarvam AI

    Sarvam 105B

    Released Feb 18, 2026

    53/100
    • ECI—
    • Price$0.158 / $0.632
    • Context131K
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (Mistral Small 4 67/100, Nemotron 3 Super 120B A12B 56/100, Sarvam 105B 53/100) 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 priceNemotron 3 Super 120B A12BNemotron 3 Super 120B A12B $0.258 · Mistral Small 4 $0.263 · Sarvam 105B $0.276 per 1M tokens (3:1 blend)
  • Longest contextNemotron 3 Super 120B A12BNemotron 3 Super 120B A12B 262,144 · Mistral Small 4 256,000 · Sarvam 105B 131,072 tokens
  • Widest inputsMistral Small 4Nemotron 3 Super 120B A12B: Text · Mistral Small 4: Text, Images · Sarvam 105B: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightNemotron 3 Super 120B A12BMistral Small 4Sarvam 105B
Price50%747473
Inputs & features30%357035
Context window20%434332
Specs & price only100%56/10067/10053/100

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.

Nemotron 3 Super 120B A12B vs Mistral Small 4 vs Sarvam 105B specifications side by side
SpecificationNemotron 3 Super 120B A12BNVIDIAMistral Small 4Mistral AISarvam 105BSarvam AI
Capability
Capabilities Index (ECI)———
ECI rank———
Price per million tokens
Input$0.135 (best)$0.15$0.158
Output$0.625$0.60 (best)$0.632
Cached input—$0.015 (best)$0.053
Blended (3:1)$0.258 (best)$0.263$0.276
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 16 providersOfficial Mistral APIOfficial Sarvam AI API
Limits
Context window262,144 tokens (best)256,000 tokens131,072 tokens
Max output262,144 tokens (best)256,000 tokens131,072 tokens
Inputs and features
TextYesYesYes
ImagesNoYesNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningYesYeshighYeslow · medium · high
Tool callingYesYesYes
Structured outputNoYesNo
Availability
WeightsOpenOpenOpen
API model ID—mistral-small-2603sarvam-105b
API providers1820 (best)3
ReleasedMar 11, 2026Mar 16, 2026Feb 18, 2026
Knowledge cutoffJun 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.

  • Nemotron 3 Super 120B A12B$2.60
  • Mistral Small 4$2.70
  • Sarvam 105B$2.84
04 — Questions

Which should you choose?

Which is better: Nemotron 3 Super 120B A12B, Mistral Small 4 or Sarvam 105B?

There is not enough shared capability evidence to name an overall winner. The scores (Mistral Small 4 67/100, Nemotron 3 Super 120B A12B 56/100, Sarvam 105B 53/100) 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, Nemotron 3 Super 120B A12B, Mistral Small 4 or Sarvam 105B?

Nemotron 3 Super 120B A12B is cheaper at $0.135 input / $0.625 output per million tokens (median across 16 API providers). Mistral Small 4 costs $0.15 input / $0.60 output per million tokens (official Mistral API price); Sarvam 105B costs $0.158 input / $0.632 output per million tokens (official Sarvam AI API price). At a typical mix of three input tokens to one output token, that is $0.258 per million tokens for Nemotron 3 Super 120B A12B versus $0.263 for Mistral Small 4 (1× as much) and $0.276 for Sarvam 105B (1.1× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. Nemotron 3 Super 120B A12B has not been scored yet, Mistral Small 4 has not been scored yet and Sarvam 105B has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for Nemotron 3 Super 120B A12B, Mistral Small 4 and Sarvam 105B 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?

Nemotron 3 Super 120B A12B has the largest context window at 262,144 tokens, against 256,000 for Mistral Small 4 and 131,072 for Sarvam 105B. Maximum output per response: Nemotron 3 Super 120B A12B up to 262,144, Mistral Small 4 up to 256,000, Sarvam 105B up to 131,072 tokens.

Which can read images, PDFs, audio or video?

Nemotron 3 Super 120B A12B accepts text; Mistral Small 4 accepts text and images; Sarvam 105B accepts text. Mistral Small 4 handles the widest range of inputs.

Are any of these open source?

Yes, all three publish their weights, so you can self-host them.

Which is newer?

Mistral Small 4 is the newest, released Mar 16, 2026. Nemotron 3 Super 120B A12B came out Mar 11, 2026; Sarvam 105B came out Feb 18, 2026. Knowledge cutoff: Nemotron 3 Super 120B A12B Jun 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.