ZMIME
Comparison · 3 models · Updated Oct 4, 2026

GPT OSS 120B vs Qwen3 235B-A22B Instruct 2507 vs DeepSeek-V3.1

GPT OSS 120B comes out ahead, 61 to 58 and 55 on our weighted score, and it is the cheaper option too.

  1. Our pick

    OpenAI

    GPT OSS 120B

    Released Aug 5, 2025

    61/100
    • ECI140.0
    • Price$0.15 / $0.60
    • Context131K
    Full specs Remove GPT OSS 120B

  2. Alibaba (Qwen)

    Qwen3 235B-A22B Instruct 2507

    Released Jul 21, 2025

    58/100
    • ECI138.9
    • Price$0.15 / $0.75
    • Context262K
    Full specs Remove Qwen3 235B-A22B Instruct 2507

  3. DeepSeek

    DeepSeek-V3.1

    Released Aug 21, 2025

    55/100
    • ECI139.9
    • Price$0.385 / $1.25
    • Context131K
    Full specs Remove DeepSeek-V3.1
01 — Verdict

GPT OSS 120B is our pick

GPT OSS 120B is the better all-round choice, scoring 61/100 against Qwen3 235B-A22B Instruct 2507 (58) and DeepSeek-V3.1 (55). It leads on price and inputs & features. Qwen3 235B-A22B Instruct 2507 wins on context window. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

  • CapabilityGPT OSS 120BCapabilities Index (ECI): GPT OSS 120B 140.0 · DeepSeek-V3.1 139.9 · Qwen3 235B-A22B Instruct 2507 138.9
  • Lowest priceGPT OSS 120BGPT OSS 120B $0.263 · Qwen3 235B-A22B Instruct 2507 $0.30 · DeepSeek-V3.1 $0.601 per 1M tokens (3:1 blend)
  • Longest contextQwen3 235B-A22B Instruct 2507Qwen3 235B-A22B Instruct 2507 262,144 · GPT OSS 120B 131,072 · DeepSeek-V3.1 131,072 tokens
  • Widest inputsSame inputsGPT OSS 120B: Text · Qwen3 235B-A22B Instruct 2507: Text · DeepSeek-V3.1: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightGPT OSS 120BQwen3 235B-A22B Instruct 2507DeepSeek-V3.1
CapabilityCapabilities Index (ECI)50%656465
Price25%777560
Inputs & features15%452535
Context window10%243724
Overall100%61/10058/10055/100
02 — Side by side

Every spec in one table

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

GPT OSS 120B vs Qwen3 235B-A22B Instruct 2507 vs DeepSeek-V3.1 specifications side by side
SpecificationGPT OSS 120BOpenAIQwen3 235B-A22B Instruct 2507Alibaba (Qwen)DeepSeek-V3.1DeepSeek
Capability
Capabilities Index (ECI)140.0 (best)138.9139.9
ECI rank#99 of 148 (best)#105 of 148#100 of 148
GPQA DiamondGraduate-level science questions75.8%——
OTIS Mock AIME 2024–2025Competition mathematics88.9%——
Price per million tokens
Input$0.15 (best)$0.15 (best)$0.385
Output$0.60 (best)$0.75$1.25
Cached input———
Blended (3:1)$0.263 (best)$0.30$0.601
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 36 providersMedian of 11 providersMedian of 8 providers
Limits
Context window131,072 tokens262,144 tokens (best)131,072 tokens
Max output32,768 tokens (best)16,384 tokens8,192 tokens
Inputs and features
TextYesYesYes
ImagesNoNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningYesNoYes
Tool callingYesYesYes
Structured outputYesNoNo
Availability
WeightsOpenOpenApache 2.0OpenMIT License
API model ID———
API providers39 (best)118
ReleasedAug 5, 2025Jul 21, 2025Aug 21, 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.

  • GPT OSS 120B$2.70
  • Qwen3 235B-A22B Instruct 2507$3.00
  • DeepSeek-V3.1$6.35
04 — Questions

Which should you choose?

Which is better: GPT OSS 120B, Qwen3 235B-A22B Instruct 2507 or DeepSeek-V3.1?

GPT OSS 120B is the better all-round choice, scoring 61/100 against Qwen3 235B-A22B Instruct 2507 (58) and DeepSeek-V3.1 (55). It leads on price and inputs & features. Qwen3 235B-A22B Instruct 2507 wins on context window. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

Which is cheaper, GPT OSS 120B, Qwen3 235B-A22B Instruct 2507 or DeepSeek-V3.1?

GPT OSS 120B is cheaper at $0.15 input / $0.60 output per million tokens (median across 36 API providers). Qwen3 235B-A22B Instruct 2507 costs $0.15 input / $0.75 output per million tokens (median across 11 API providers); DeepSeek-V3.1 costs $0.385 input / $1.25 output per million tokens (median across 8 API providers). At a typical mix of three input tokens to one output token, that is $0.263 per million tokens for GPT OSS 120B versus $0.30 for Qwen3 235B-A22B Instruct 2507 (1.1× as much) and $0.601 for DeepSeek-V3.1 (2.3× as much).

Which scores higher on benchmarks?

GPT OSS 120B scores higher on the Capabilities Index (ECI): GPT OSS 120B 140.0 (#99 of 148), DeepSeek-V3.1 139.9 (#100 of 148) and Qwen3 235B-A22B Instruct 2507 138.9 (#105 of 148). The confidence ranges of the top two overlap (135.3–142.3 vs 136.1–143.3), so treat the gap as small.

Which is better for coding?

There are no published SWE-bench Verified results for GPT OSS 120B, Qwen3 235B-A22B Instruct 2507 and DeepSeek-V3.1 yet, so there is no like-for-like coding score. On overall capability, GPT OSS 120B 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?

Qwen3 235B-A22B Instruct 2507 has the largest context window at 262,144 tokens, against 131,072 for GPT OSS 120B and 131,072 for DeepSeek-V3.1. Maximum output per response: GPT OSS 120B up to 32,768, Qwen3 235B-A22B Instruct 2507 up to 16,384, DeepSeek-V3.1 up to 8,192 tokens.

Which can read images, PDFs, audio or video?

GPT OSS 120B accepts text; Qwen3 235B-A22B Instruct 2507 accepts text; DeepSeek-V3.1 accepts text. They handle the same number of input types.

Are any of these open source?

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

Which is newer?

DeepSeek-V3.1 is the newest, released Aug 21, 2025. GPT OSS 120B came out Aug 5, 2025; Qwen3 235B-A22B Instruct 2507 came out Jul 21, 2025.

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