ZMIME
Comparison · 3 models · Updated Nov 18, 2026

DeepSeek V3.2 vs GPT-5.4 nano vs Qwen3.6 27B

GPT-5.4 nano comes out ahead, 68 to 64 and 63 on our weighted score, though DeepSeek V3.2 is 32% cheaper per token.

Price in rating: Included · Exclude price

  1. DeepSeek

    DeepSeek V3.2

    Released Dec 1, 2025

    63/100
    • ECI146.3
    • Price$0.28 / $0.42
    • Context128K
  2. Our pick

    OpenAI

    GPT-5.4 nano

    Released Mar 17, 2026

    68/100
    • ECI146.3
    • Price$0.20 / $1.25
    • Context400K
  3. Alibaba (Qwen)

    Qwen3.6 27B

    Released Apr 22, 2026

    64/100
    • ECI146.8
    • Price$0.60 / $3.60
    • Context262K
01 — Verdict

GPT-5.4 nano is our pick

GPT-5.4 nano is the better all-round choice, scoring 68/100 against Qwen3.6 27B (64) and DeepSeek V3.2 (63). It leads on context window. DeepSeek V3.2 wins on price. Qwen3.6 27B wins on inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

  • CapabilityQwen3.6 27BCapabilities Index (ECI): Qwen3.6 27B 146.8 · GPT-5.4 nano 146.3 · DeepSeek V3.2 146.3
  • Lowest priceDeepSeek V3.2DeepSeek V3.2 $0.315 · GPT-5.4 nano $0.463 · Qwen3.6 27B $1.35 per 1M tokens (3:1 blend)
  • Longest contextGPT-5.4 nanoGPT-5.4 nano 400,000 · Qwen3.6 27B 262,144 · DeepSeek V3.2 128,000 tokens
  • Widest inputsQwen3.6 27BDeepSeek V3.2: Text · GPT-5.4 nano: Text, Images · Qwen3.6 27B: Text, Images, Video
  • Self-hostingDeepSeek V3.2 and Qwen3.6 27BPublishes downloadable weights (MIT License)
How the score is built
MeasureWeightDeepSeek V3.2GPT-5.4 nanoQwen3.6 27B
CapabilityCapabilities Index (ECI)50%747474
Price25%716342
Inputs & features15%357080
Context window10%325043
Overall100%63/10068/10064/100
02 — Side by side

Every spec in one table

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

DeepSeek V3.2 vs GPT-5.4 nano vs Qwen3.6 27B specifications side by side
SpecificationDeepSeek V3.2DeepSeekGPT-5.4 nanoOpenAIQwen3.6 27BAlibaba (Qwen)
Capability
Capabilities Index (ECI)146.3146.3146.8 (best)
ECI rank#87 of 158#86 of 158#84 of 158 (best)
GPQA DiamondGraduate-level science questions83.4%78.5%85.9% (best)
FrontierMath Tiers 1–3Research-level mathematics—44.9% (best)35.1%
OTIS Mock AIME 2024–2025Competition mathematics87.8%87.8%91.1% (best)
SimpleQA VerifiedShort factual questions—11.7%—
Price per million tokens
Input$0.28$0.20 (best)$0.60
Output$0.42 (best)$1.25$3.60
Cached input—$0.02—
Blended (3:1)$0.315 (best)$0.463$1.35
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 30 providersOfficial OpenAI APIOfficial Alibaba API
Limits
Context window128,000 tokens400,000 tokens (best)262,144 tokens
Max output64,000 tokens128,000 tokens (best)65,536 tokens
Inputs and features
TextYesYesYes
ImagesNoYesYes
PDFsNoNoNo
AudioNoNoNo
VideoNoNoYes
ReasoningYesYeslow · medium · high · xhighYes
Tool callingYesYesYes
Structured outputNoYesYes
Availability
WeightsOpenMIT LicenseProprietaryOpen
API model ID—gpt-5.4-nanoqwen3.6-27b
API providers3036 (best)29
ReleasedDec 1, 2025Mar 17, 2026Apr 22, 2026
Knowledge cutoffJul 2024Aug 31, 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.

  • DeepSeek V3.2$3.64
  • GPT-5.4 nano$4.50
  • Qwen3.6 27B$13.20
04 — Questions

Which should you choose?

Which is better: DeepSeek V3.2, GPT-5.4 nano or Qwen3.6 27B?

GPT-5.4 nano is the better all-round choice, scoring 68/100 against Qwen3.6 27B (64) and DeepSeek V3.2 (63). It leads on context window. DeepSeek V3.2 wins on price. Qwen3.6 27B wins on inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

Which is cheaper, DeepSeek V3.2, GPT-5.4 nano or Qwen3.6 27B?

DeepSeek V3.2 is cheaper at $0.28 input / $0.42 output per million tokens (median across 30 API providers). GPT-5.4 nano costs $0.20 input / $1.25 output per million tokens (official OpenAI API price); Qwen3.6 27B costs $0.60 input / $3.60 output per million tokens (official Alibaba API price). At a typical mix of three input tokens to one output token, that is $0.315 per million tokens for DeepSeek V3.2 versus $0.463 for GPT-5.4 nano (1.5× as much) and $1.35 for Qwen3.6 27B (4.3× as much).

Which scores higher on benchmarks?

Qwen3.6 27B scores higher on the Capabilities Index (ECI): Qwen3.6 27B 146.8 (#84 of 158), GPT-5.4 nano 146.3 (#86 of 158) and DeepSeek V3.2 146.3 (#87 of 158). The confidence ranges of the top two overlap (144.7–148.9 vs 144.7–148.0), so treat the gap as small. On individual benchmarks: GPQA Diamond — Qwen3.6 27B 85.9%, DeepSeek V3.2 83.4%, GPT-5.4 nano 78.5%; OTIS Mock AIME 2024–2025 — Qwen3.6 27B 91.1%, DeepSeek V3.2 87.8%, GPT-5.4 nano 87.8%.

Which is better for coding?

There are no published SWE-bench Verified results for DeepSeek V3.2, GPT-5.4 nano and Qwen3.6 27B yet, so there is no like-for-like coding score. On overall capability, Qwen3.6 27B 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?

GPT-5.4 nano has the largest context window at 400,000 tokens, against 262,144 for Qwen3.6 27B and 128,000 for DeepSeek V3.2. Maximum output per response: DeepSeek V3.2 up to 64,000, GPT-5.4 nano up to 128,000, Qwen3.6 27B up to 65,536 tokens.

Which can read images, PDFs, audio or video?

DeepSeek V3.2 accepts text; GPT-5.4 nano accepts text and images; Qwen3.6 27B accepts text, images and video. Qwen3.6 27B handles the widest range of inputs.

Are any of these open source?

DeepSeek V3.2 and Qwen3.6 27B publishes its weights (MIT License) and can be self-hosted; GPT-5.4 nano is proprietary.

Which is newer?

Qwen3.6 27B is the newest, released Apr 22, 2026. GPT-5.4 nano came out Mar 17, 2026; DeepSeek V3.2 came out Dec 1, 2025. Knowledge cutoff: DeepSeek V3.2 Jul 2024, GPT-5.4 nano Aug 31, 2025.

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.