ZMIME
Comparison · 3 models · Updated Oct 4, 2026

DeepSeek-R1-Distill-Qwen-32B vs GPT OSS 20B vs QwQ 32B

Too close to call on our weighted score (GPT OSS 20B 54, QwQ 32B 52, DeepSeek-R1-Distill-Qwen-32B 47). The right pick depends on what you value most.

  1. DeepSeek

    DeepSeek-R1-Distill-Qwen-32B

    Released Jan 20, 2025

    47/100
    • ECI137.4
    • Price—
    • Context131K
    Full specs Remove DeepSeek-R1-Distill-Qwen-32B

  2. OpenAI

    GPT OSS 20B

    Released Aug 5, 2025

    54/100
    • ECI137.8
    • Price$0.07 / $0.295
    • Context131K
    Full specs Remove GPT OSS 20B

  3. Alibaba (Qwen)

    QwQ 32B

    Released Mar 5, 2025

    52/100
    • ECI137.6
    • Price$0.66 / $1.00
    • Context131K
    Full specs Remove QwQ 32B
01 — Verdict

Too close to call

It is close. Our weighted score puts them within 2 points (GPT OSS 20B 54/100, QwQ 32B 52/100, DeepSeek-R1-Distill-Qwen-32B 47/100), so choose by what matters most for your work: GPT OSS 20B for raw capability. The score weighs capability 67%, inputs & features 20%, context window 13%.

  • CapabilityGPT OSS 20BCapabilities Index (ECI): GPT OSS 20B 137.8 · QwQ 32B 137.6 · DeepSeek-R1-Distill-Qwen-32B 137.4
  • Lowest priceGPT OSS 20BGPT OSS 20B $0.126 · QwQ 32B $0.745 per 1M tokens (3:1 blend) · DeepSeek-R1-Distill-Qwen-32B unpriced
  • Longest contextAbout the sameDeepSeek-R1-Distill-Qwen-32B 131,072 · GPT OSS 20B 131,072 · QwQ 32B 131,072 tokens
  • Widest inputsSame inputsDeepSeek-R1-Distill-Qwen-32B: Text · GPT OSS 20B: Text · QwQ 32B: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightDeepSeek-R1-Distill-Qwen-32BGPT OSS 20BQwQ 32B
CapabilityCapabilities Index (ECI)67%626362
Inputs & features20%104535
Context window13%242424
Overall100%47/10054/10052/100

Left out because at least one model lacks the data: price. 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.

DeepSeek-R1-Distill-Qwen-32B vs GPT OSS 20B vs QwQ 32B specifications side by side
SpecificationDeepSeek-R1-Distill-Qwen-32BDeepSeekGPT OSS 20BOpenAIQwQ 32BAlibaba (Qwen)
Capability
Capabilities Index (ECI)137.4137.8 (best)137.6
ECI rank#110 of 148#108 of 148 (best)#109 of 148
GPQA DiamondGraduate-level science questions64.1%60.8%65.3% (best)
OTIS Mock AIME 2024–2025Competition mathematics55.6%65.3% (best)59.2%
Price per million tokens
Input—$0.07 (best)$0.66
Output—$0.295 (best)$1.00
Cached input———
Blended (3:1)—$0.126 (best)$0.745
Long-context rate—Same rateSame rate
Price source—Median of 18 providersMedian of 1 providers
Limits
Context window131,072 tokens131,072 tokens131,072 tokens
Max output32,768 tokens (best)32,768 tokens (best)8,192 tokens
Inputs and features
TextYesYesYes
ImagesNoNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningYesYesYes
Tool callingNoYesYes
Structured outputNoYesNo
Availability
WeightsOpenOpenOpen
API model ID———
API providers—19 (best)1
ReleasedJan 20, 2025Aug 5, 2025Mar 5, 2025
Knowledge cutoff——Apr 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-R1-Distill-Qwen-32B—
  • GPT OSS 20B$1.29
  • QwQ 32B$8.60
04 — Questions

Which should you choose?

Which is better: DeepSeek-R1-Distill-Qwen-32B, GPT OSS 20B or QwQ 32B?

It is close. Our weighted score puts them within 2 points (GPT OSS 20B 54/100, QwQ 32B 52/100, DeepSeek-R1-Distill-Qwen-32B 47/100), so choose by what matters most for your work: GPT OSS 20B for raw capability. The score weighs capability 67%, inputs & features 20%, context window 13%.

Which is cheaper, DeepSeek-R1-Distill-Qwen-32B, GPT OSS 20B or QwQ 32B?

GPT OSS 20B is cheaper at $0.07 input / $0.295 output per million tokens (median across 18 API providers). QwQ 32B costs $0.66 input / $1.00 output per million tokens (median across 1 API provider). At a typical mix of three input tokens to one output token, that is $0.126 per million tokens for GPT OSS 20B versus $0.745 for QwQ 32B (5.9× as much). DeepSeek-R1-Distill-Qwen-32B has no published per-token price.

Which scores higher on benchmarks?

GPT OSS 20B scores higher on the Capabilities Index (ECI): GPT OSS 20B 137.8 (#108 of 148), QwQ 32B 137.6 (#109 of 148) and DeepSeek-R1-Distill-Qwen-32B 137.4 (#110 of 148). The confidence ranges of the top two overlap (133.0–139.6 vs 133.1–141.7), so treat the gap as small. On individual benchmarks: GPQA Diamond — QwQ 32B 65.3%, DeepSeek-R1-Distill-Qwen-32B 64.1%, GPT OSS 20B 60.8%; OTIS Mock AIME 2024–2025 — GPT OSS 20B 65.3%, QwQ 32B 59.2%, DeepSeek-R1-Distill-Qwen-32B 55.6%.

Which is better for coding?

There are no published SWE-bench Verified results for DeepSeek-R1-Distill-Qwen-32B, GPT OSS 20B and QwQ 32B yet, so there is no like-for-like coding score. On overall capability, GPT OSS 20B leads, which tends to carry over to coding, but test on your own codebase. Note that DeepSeek-R1-Distill-Qwen-32B does not support tool calling, which most coding agents need.

Which has the bigger context window?

DeepSeek-R1-Distill-Qwen-32B, GPT OSS 20B and QwQ 32B share the same 131,072-token context window. Maximum output per response: DeepSeek-R1-Distill-Qwen-32B up to 32,768, GPT OSS 20B up to 32,768, QwQ 32B up to 8,192 tokens.

Which can read images, PDFs, audio or video?

DeepSeek-R1-Distill-Qwen-32B accepts text; GPT OSS 20B accepts text; QwQ 32B accepts text. They handle the same number of input types.

Are any of these open source?

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

Which is newer?

GPT OSS 20B is the newest, released Aug 5, 2025. QwQ 32B came out Mar 5, 2025; DeepSeek-R1-Distill-Qwen-32B came out Jan 20, 2025. Knowledge cutoff: QwQ 32B Apr 2024.

How do you decide the winner?

Each model gets a 0–100 score on 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 Oct 4, 2026.

05 — Keep comparing

Related comparisons