ZMIME
Comparison · 3 models · Updated Nov 18, 2026

DeepSeek-V3 vs Magistral Small vs Qwen3 8B

Qwen3 8B comes out ahead, 46 to 40 and 33 on our weighted score, and it is the cheaper option too.

Price in rating: Excluded · Include price

  1. DeepSeek

    DeepSeek-V3

    Released Dec 26, 2024

    33/100
    • ECI131.7
    • Price$0.32 / $1.10
    • Context131K
  2. Mistral AI

    Magistral Small

    Released Jun 10, 2025

    40/100
    • ECI—
    • Price$0.50 / $1.50
    • Context131K
  3. Our pick

    Alibaba (Qwen)

    Qwen3 8B

    Released Apr 29, 2025Deprecated

    46/100
    • ECI132.1
    • Price$0.18 / $0.70
    • Context33K
01 — Verdict

Qwen3 8B is our pick

Qwen3 8B is the better all-round choice, scoring 46/100 against Magistral Small (40) and DeepSeek-V3 (33). It leads on capability. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond and OTIS Mock AIME 2024–2025), because Magistral Small has no Capabilities Index score yet.

  • CapabilityQwen3 8BShared benchmarks: Qwen3 8B 56.4% · Magistral Small 43.0% · DeepSeek-V3 36.2%
  • Lowest priceQwen3 8BQwen3 8B $0.31 · DeepSeek-V3 $0.515 · Magistral Small $0.75 per 1M tokens (3:1 blend)
  • Longest contextDeepSeek-V3 and Magistral SmallDeepSeek-V3 131,072 · Magistral Small 131,072 · Qwen3 8B 32,768 tokens
  • Widest inputsSame inputsDeepSeek-V3: Text · Magistral Small: Text · Qwen3 8B: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightDeepSeek-V3Magistral SmallQwen3 8B
CapabilityShared benchmarks67%364356
Inputs & features20%253535
Context window13%323211
Overall100%33/10040/10046/100
02 — Side by side

Every spec in one table

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

DeepSeek-V3 vs Magistral Small vs Qwen3 8B specifications side by side
SpecificationDeepSeek-V3DeepSeekMagistral SmallMistral AIQwen3 8BAlibaba (Qwen)
Capability
Capabilities Index (ECI)131.7—132.1 (best)
ECI rank#136 of 158—#135 of 158 (best)
GPQA DiamondGraduate-level science questions56.5%56.1%56.8% (best)
OTIS Mock AIME 2024–2025Competition mathematics15.8%30.0%56.1% (best)
Price per million tokens
Input$0.32$0.50$0.18 (best)
Output$1.10$1.50$0.70 (best)
Cached input———
Blended (3:1)$0.515$0.75$0.31 (best)
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 9 providersMedian of 1 providersOfficial Alibaba API
Limits
Context window131,072 tokens (best)131,072 tokens (best)32,768 tokens
Max output8,192 tokens8,192 tokens32,768 tokens (best)
Inputs and features
TextYesYesYes
ImagesNoNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningNoYesYes
Tool callingYesYesYes
Structured outputNoNoNo
Availability
WeightsOpenDeepSeek Model LicenseOpenApache 2.0OpenApache-2.0
API model ID——qwen3-8b
API providers9 (best)26
ReleasedDec 26, 2024Jun 10, 2025Apr 29, 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.

  • DeepSeek-V3$5.40
  • Magistral Small$8.00
  • Qwen3 8B$3.20
04 — Questions

Which should you choose?

Which is better: DeepSeek-V3, Magistral Small or Qwen3 8B?

Qwen3 8B is the better all-round choice, scoring 46/100 against Magistral Small (40) and DeepSeek-V3 (33). It leads on capability. The score weighs capability 67%, inputs & features 20%, context window 13%. 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, Magistral Small or Qwen3 8B?

Qwen3 8B is cheaper at $0.18 input / $0.70 output per million tokens (official Alibaba API price). 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 $0.31 per million tokens for Qwen3 8B versus $0.515 for DeepSeek-V3 (1.7× as much) and $0.75 for Magistral Small (2.4× 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): Qwen3 8B 56.4%, Magistral Small 43.0% and DeepSeek-V3 36.2%. On individual benchmarks: GPQA Diamond — Qwen3 8B 56.8%, DeepSeek-V3 56.5%, Magistral Small 56.1%; OTIS Mock AIME 2024–2025 — Qwen3 8B 56.1%, Magistral Small 30.0%, DeepSeek-V3 15.8%.

Which is better for coding?

There are no published SWE-bench Verified results for DeepSeek-V3, Magistral Small and Qwen3 8B yet, so there is no like-for-like coding score. On overall capability, Qwen3 8B 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 and Magistral Small have the largest context windows (131,072 and 131,072 tokens), against 32,768 for Qwen3 8B. Maximum output per response: DeepSeek-V3 up to 8,192, Magistral Small up to 8,192, Qwen3 8B up to 32,768 tokens.

Which can read images, PDFs, audio or video?

DeepSeek-V3 accepts text; Magistral Small accepts text; Qwen3 8B accepts text. They handle the same number of input types.

Are any of these open source?

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

Which is newer?

Magistral Small is the newest, released Jun 10, 2025. Qwen3 8B came out Apr 29, 2025; DeepSeek-V3 came out Dec 26, 2024.

How do you decide the winner?

This rating uses capability (67%, independent benchmark results); inputs & features (20%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (13%, 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.