ZMIME
Comparison · 3 models · Updated Nov 2, 2026

Llama 4 Maverick 17B Instruct vs GPT OSS 20B vs DeepSeek-V3

GPT OSS 20B comes out ahead, 61 to 57 and 49 on our weighted score, and it is the cheaper option too.

  1. Meta

    Llama 4 Maverick 17B Instruct

    Released Apr 5, 2025

    57/100
    • ECI132.3
    • Price$0.245 / $0.838
    • Context1M
    Full specs Remove Llama 4 Maverick 17B Instruct
  2. Our pick

    OpenAI

    GPT OSS 20B

    Released Aug 5, 2025

    61/100
    • ECI137.8
    • Price$0.07 / $0.25
    • Context131K
    Full specs Remove GPT OSS 20B

  3. DeepSeek

    DeepSeek-V3

    Released Dec 26, 2024

    49/100
    • ECI132.3
    • Price$0.32 / $1.10
    • Context131K
    Full specs Remove DeepSeek-V3
01 — Verdict

GPT OSS 20B is our pick

GPT OSS 20B is the better all-round choice, scoring 61/100 against Llama 4 Maverick 17B Instruct (57) and DeepSeek-V3 (49). It leads on capability and price. Llama 4 Maverick 17B Instruct wins on inputs & features and context window. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

  • CapabilityGPT OSS 20BCapabilities Index (ECI): GPT OSS 20B 137.8 · Llama 4 Maverick 17B Instruct 132.3 · DeepSeek-V3 132.3
  • Lowest priceGPT OSS 20BGPT OSS 20B $0.115 · Llama 4 Maverick 17B Instruct $0.393 · DeepSeek-V3 $0.515 per 1M tokens (3:1 blend)
  • Longest contextLlama 4 Maverick 17B InstructLlama 4 Maverick 17B Instruct 1,000,000 · GPT OSS 20B 131,072 · DeepSeek-V3 131,072 tokens
  • Widest inputsLlama 4 Maverick 17B InstructLlama 4 Maverick 17B Instruct: Text, Images · GPT OSS 20B: Text · DeepSeek-V3: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightLlama 4 Maverick 17B InstructGPT OSS 20BDeepSeek-V3
CapabilityCapabilities Index (ECI)50%556155
Price25%669061
Inputs & features15%503525
Context window10%602525
Overall100%57/10061/10049/100
02 — Side by side

Every spec in one table

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

Llama 4 Maverick 17B Instruct vs GPT OSS 20B vs DeepSeek-V3 specifications side by side
SpecificationLlama 4 Maverick 17B InstructMetaGPT OSS 20BOpenAIDeepSeek-V3DeepSeek
Capability
Capabilities Index (ECI)132.3137.8 (best)132.3
ECI rank#135 of 165#122 of 165 (best)#136 of 165
GPQA DiamondGraduate-level science questions67.0% (best)60.8%56.5%
OTIS Mock AIME 2024–2025Competition mathematics20.6%65.3% (best)15.8%
Price per million tokens
Input$0.245$0.07 (best)$0.32
Output$0.838$0.25 (best)$1.10
Cached input———
Blended (3:1)$0.393$0.115 (best)$0.515
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 16 providersMedian of 37 providersMedian of 8 providers
Limits
Context window1,000,000 tokens (best)131,072 tokens131,072 tokens
Max output16,384 tokens32,768 tokens (best)8,192 tokens
Inputs and features
TextYesYesYes
ImagesYesNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningNoYesNo
Tool callingYesYesYes
Structured outputNoNoNo
Availability
WeightsOpenOpenOpenDeepSeek Model License
API model ID———
API providers1641 (best)8
ReleasedApr 5, 2025Aug 5, 2025Dec 26, 2024
Knowledge cutoffAug 2024Jun 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.

  • Llama 4 Maverick 17B Instruct$4.13
  • GPT OSS 20B$1.20
  • DeepSeek-V3$5.40
04 — Questions

Which should you choose?

Which is better: Llama 4 Maverick 17B Instruct, GPT OSS 20B or DeepSeek-V3?

GPT OSS 20B is the better all-round choice, scoring 61/100 against Llama 4 Maverick 17B Instruct (57) and DeepSeek-V3 (49). It leads on capability and price. Llama 4 Maverick 17B Instruct wins on inputs & features and context window. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

Which is cheaper, Llama 4 Maverick 17B Instruct, GPT OSS 20B or DeepSeek-V3?

GPT OSS 20B is cheaper at $0.07 input / $0.25 output per million tokens (median across 37 API providers). Llama 4 Maverick 17B Instruct costs $0.245 input / $0.838 output per million tokens (median across 16 API providers); DeepSeek-V3 costs $0.32 input / $1.10 output per million tokens (median across 8 API providers). At a typical mix of three input tokens to one output token, that is $0.115 per million tokens for GPT OSS 20B versus $0.393 for Llama 4 Maverick 17B Instruct (3.4× as much) and $0.515 for DeepSeek-V3 (4.5× as much).

Which scores higher on benchmarks?

GPT OSS 20B scores higher on the Capabilities Index (ECI): GPT OSS 20B 137.8 (#122 of 165), Llama 4 Maverick 17B Instruct 132.3 (#135 of 165) and DeepSeek-V3 132.3 (#136 of 165). The confidence ranges of the top two overlap (133.1–139.7 vs 128.6–134.2), so treat the gap as small. On individual benchmarks: GPQA Diamond — Llama 4 Maverick 17B Instruct 67.0%, GPT OSS 20B 60.8%, DeepSeek-V3 56.5%; OTIS Mock AIME 2024–2025 — GPT OSS 20B 65.3%, Llama 4 Maverick 17B Instruct 20.6%, DeepSeek-V3 15.8%.

Which is better for coding?

There are no published SWE-bench Verified results for Llama 4 Maverick 17B Instruct, GPT OSS 20B and DeepSeek-V3 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. All three support tool calling for agent workflows.

Which has the bigger context window?

Llama 4 Maverick 17B Instruct has the largest context window at 1,000,000 tokens, against 131,072 for GPT OSS 20B and 131,072 for DeepSeek-V3. Maximum output per response: Llama 4 Maverick 17B Instruct up to 16,384, GPT OSS 20B up to 32,768, DeepSeek-V3 up to 8,192 tokens.

Which can read images, PDFs, audio or video?

Llama 4 Maverick 17B Instruct accepts text and images; GPT OSS 20B accepts text; DeepSeek-V3 accepts text. Llama 4 Maverick 17B Instruct handles the widest range of inputs.

Are any of these open source?

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

Which is newer?

GPT OSS 20B is the newest, released Aug 5, 2025. Llama 4 Maverick 17B Instruct came out Apr 5, 2025; DeepSeek-V3 came out Dec 26, 2024. Knowledge cutoff: Llama 4 Maverick 17B Instruct Aug 2024, GPT OSS 20B Jun 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 Nov 2, 2026.

05 — Keep comparing

Related comparisons