ZMIME
Comparison · 3 models · Updated Nov 18, 2026

DeepSeek-V3 vs Gemma 3 27B IT vs Magistral Small

Gemma 3 27B IT comes out ahead, 50 to 43 and 40 on our weighted score, and it is the cheaper option too.

Price in rating: Included · Exclude price

  1. DeepSeek

    DeepSeek-V3

    Released Dec 26, 2024

    40/100
    • ECI131.7
    • Price$0.32 / $1.10
    • Context131K
  2. Our pick

    Google

    Gemma 3 27B IT

    Released Mar 12, 2025

    50/100
    • ECI127.5
    • Price$0.10 / $0.299
    • Context131K
  3. Mistral AI

    Magistral Small

    Released Jun 10, 2025

    43/100
    • ECI—
    • Price$0.50 / $1.50
    • Context131K
01 — Verdict

Gemma 3 27B IT is our pick

Gemma 3 27B IT is the better all-round choice, scoring 50/100 against Magistral Small (43) and DeepSeek-V3 (40). It leads on price and inputs & features. Magistral Small wins on capability. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%. Capability uses the average of shared benchmarks (GPQA Diamond and OTIS Mock AIME 2024–2025), because Magistral Small has no Capabilities Index score yet.

  • CapabilityMagistral SmallShared benchmarks: Magistral Small 43.0% · DeepSeek-V3 36.2% · Gemma 3 27B IT 35.1%
  • Lowest priceGemma 3 27B ITGemma 3 27B IT $0.15 · DeepSeek-V3 $0.515 · Magistral Small $0.75 per 1M tokens (3:1 blend)
  • Longest contextAbout the sameDeepSeek-V3 131,072 · Gemma 3 27B IT 131,072 · Magistral Small 131,072 tokens
  • Widest inputsGemma 3 27B ITDeepSeek-V3: Text · Gemma 3 27B IT: Text, Images · Magistral Small: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightDeepSeek-V3Gemma 3 27B ITMagistral Small
CapabilityShared benchmarks50%363543
Price25%618553
Inputs & features15%255035
Context window10%323232
Overall100%40/10050/10043/100
02 — Side by side

Every spec in one table

Highlighted cells lead their row. Dashes mean the data is not published.

DeepSeek-V3 vs Gemma 3 27B IT vs Magistral Small specifications side by side
SpecificationDeepSeek-V3DeepSeekGemma 3 27B ITGoogleMagistral SmallMistral AI
Capability
Capabilities Index (ECI)131.7 (best)127.5—
ECI rank#136 of 158 (best)#141 of 158—
GPQA DiamondGraduate-level science questions56.5% (best)47.7%56.1%
OTIS Mock AIME 2024–2025Competition mathematics15.8%22.5%30.0% (best)
Price per million tokens
Input$0.32$0.10 (best)$0.50
Output$1.10$0.299 (best)$1.50
Cached input———
Blended (3:1)$0.515$0.15 (best)$0.75
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 9 providersMedian of 17 providersMedian of 1 providers
Limits
Context window131,072 tokens131,072 tokens131,072 tokens
Max output8,192 tokens131,072 tokens (best)8,192 tokens
Inputs and features
TextYesYesYes
ImagesNoYesNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningNoNoYes
Tool callingYesYesYes
Structured outputNoNoNo
Availability
WeightsOpenDeepSeek Model LicenseOpenOpenApache 2.0
API model ID———
API providers918 (best)2
ReleasedDec 26, 2024Mar 12, 2025Jun 10, 2025
Knowledge cutoff—Aug 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.

  • DeepSeek-V3$5.40
  • Gemma 3 27B IT$1.60
  • Magistral Small$8.00
04 — Questions

Which should you choose?

Which is better: DeepSeek-V3, Gemma 3 27B IT or Magistral Small?

Gemma 3 27B IT is the better all-round choice, scoring 50/100 against Magistral Small (43) and DeepSeek-V3 (40). It leads on price and inputs & features. Magistral Small wins on capability. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%. Capability uses the average of shared benchmarks (GPQA Diamond and OTIS Mock AIME 2024–2025), because Magistral Small has no Capabilities Index score yet.

Which is cheaper, DeepSeek-V3, Gemma 3 27B IT or Magistral Small?

Gemma 3 27B IT is cheaper at $0.10 input / $0.299 output per million tokens (median across 17 API providers). DeepSeek-V3 costs $0.32 input / $1.10 output per million tokens (median across 9 API providers); Magistral Small costs $0.50 input / $1.50 output per million tokens (median across 1 API provider). At a typical mix of three input tokens to one output token, that is $0.15 per million tokens for Gemma 3 27B IT versus $0.515 for DeepSeek-V3 (3.4× as much) and $0.75 for Magistral Small (5× as much).

Which scores higher on benchmarks?

Not every model here has a Capabilities Index score, so we compare the average of shared benchmarks (GPQA Diamond and OTIS Mock AIME 2024–2025): Magistral Small 43.0%, DeepSeek-V3 36.2% and Gemma 3 27B IT 35.1%. On individual benchmarks: GPQA Diamond — DeepSeek-V3 56.5%, Magistral Small 56.1%, Gemma 3 27B IT 47.7%; OTIS Mock AIME 2024–2025 — Magistral Small 30.0%, Gemma 3 27B IT 22.5%, DeepSeek-V3 15.8%.

Which is better for coding?

There are no published SWE-bench Verified results for DeepSeek-V3, Gemma 3 27B IT and Magistral Small yet, so there is no like-for-like coding score. On overall capability, Magistral Small leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.

Which has the bigger context window?

DeepSeek-V3, Gemma 3 27B IT and Magistral Small share the same 131,072-token context window. Maximum output per response: DeepSeek-V3 up to 8,192, Gemma 3 27B IT up to 131,072, Magistral Small up to 8,192 tokens.

Which can read images, PDFs, audio or video?

DeepSeek-V3 accepts text; Gemma 3 27B IT accepts text and images; Magistral Small accepts text. Gemma 3 27B IT handles the widest range of inputs.

Are any of these open source?

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

Which is newer?

Magistral Small is the newest, released Jun 10, 2025. Gemma 3 27B IT came out Mar 12, 2025; DeepSeek-V3 came out Dec 26, 2024. Knowledge cutoff: Gemma 3 27B IT Aug 2024.

How do you decide the winner?

This rating uses capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, 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.