ZMIME
Comparison · 3 models · Updated Nov 2, 2026

Gemma 4 31B IT vs GPT OSS 120B vs Qwen3 235B-A22B Instruct 2507

Gemma 4 31B IT comes out ahead, 68 to 58 and 57 on our weighted score, and it is the cheaper option too.

  1. Our pick

    Google

    Gemma 4 31B IT

    Released Apr 2, 2026

    68/100
    • ECI143.1
    • Price$0.14 / $0.40
    • Context262K
    Full specs Remove Gemma 4 31B IT

  2. OpenAI

    GPT OSS 120B

    Released Aug 5, 2025

    58/100
    • ECI139.8
    • Price$0.15 / $0.60
    • Context131K
    Full specs Remove GPT OSS 120B

  3. Alibaba (Qwen)

    Qwen3 235B-A22B Instruct 2507

    Released Jul 21, 2025

    57/100
    • ECI138.8
    • Price$0.199 / $0.60
    • Context262K
    Full specs Remove Qwen3 235B-A22B Instruct 2507
01 — Verdict

Gemma 4 31B IT is our pick

Gemma 4 31B IT is the better all-round choice, scoring 68/100 against GPT OSS 120B (58) and Qwen3 235B-A22B Instruct 2507 (57). It leads on capability, price and inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

  • CapabilityGemma 4 31B ITCapabilities Index (ECI): Gemma 4 31B IT 143.1 · GPT OSS 120B 139.8 · Qwen3 235B-A22B Instruct 2507 138.8
  • Lowest priceGemma 4 31B ITGemma 4 31B IT $0.205 · GPT OSS 120B $0.263 · Qwen3 235B-A22B Instruct 2507 $0.299 per 1M tokens (3:1 blend)
  • Longest contextGemma 4 31B IT and Qwen3 235B-A22B Instruct 2507Gemma 4 31B IT 262,144 · Qwen3 235B-A22B Instruct 2507 262,144 · GPT OSS 120B 131,072 tokens
  • Widest inputsGemma 4 31B ITGemma 4 31B IT: Text, Images · GPT OSS 120B: Text · Qwen3 235B-A22B Instruct 2507: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightGemma 4 31B ITGPT OSS 120BQwen3 235B-A22B Instruct 2507
CapabilityCapabilities Index (ECI)50%686463
Price25%797472
Inputs & features15%703525
Context window10%372537
Overall100%68/10058/10057/100
02 — Side by side

Every spec in one table

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

Gemma 4 31B IT vs GPT OSS 120B vs Qwen3 235B-A22B Instruct 2507 specifications side by side
SpecificationGemma 4 31B ITGoogleGPT OSS 120BOpenAIQwen3 235B-A22B Instruct 2507Alibaba (Qwen)
Capability
Capabilities Index (ECI)143.1 (best)139.8138.8
ECI rank#96 of 165 (best)#114 of 165#118 of 165
GPQA DiamondGraduate-level science questions75.8%75.8%—
OTIS Mock AIME 2024–2025Competition mathematics73.3%88.9% (best)—
SimpleQA VerifiedShort factual questions10.4%——
Price per million tokens
Input$0.14 (best)$0.15$0.199
Output$0.40 (best)$0.60$0.60
Cached input———
Blended (3:1)$0.205 (best)$0.263$0.299
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 36 providersMedian of 65 providersMedian of 23 providers
Limits
Context window262,144 tokens (best)131,072 tokens262,144 tokens (best)
Max output32,768 tokens (best)32,768 tokens (best)16,384 tokens
Inputs and features
TextYesYesYes
ImagesYesNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningYesYesNo
Tool callingYesYesYes
Structured outputYesNoNo
Availability
WeightsOpenOpenOpenApache 2.0
API model IDgemma-4-31b-it——
API providers4667 (best)25
ReleasedApr 2, 2026Aug 5, 2025Jul 21, 2025
Knowledge cutoffJan 2025Jun 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.

  • Gemma 4 31B IT$2.20
  • GPT OSS 120B$2.70
  • Qwen3 235B-A22B Instruct 2507$3.19
04 — Questions

Which should you choose?

Which is better: Gemma 4 31B IT, GPT OSS 120B or Qwen3 235B-A22B Instruct 2507?

Gemma 4 31B IT is the better all-round choice, scoring 68/100 against GPT OSS 120B (58) and Qwen3 235B-A22B Instruct 2507 (57). It leads on capability, price and inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

Which is cheaper, Gemma 4 31B IT, GPT OSS 120B or Qwen3 235B-A22B Instruct 2507?

Gemma 4 31B IT is cheaper at $0.14 input / $0.40 output per million tokens (median across 36 API providers). GPT OSS 120B costs $0.15 input / $0.60 output per million tokens (median across 65 API providers); Qwen3 235B-A22B Instruct 2507 costs $0.199 input / $0.60 output per million tokens (median across 23 API providers). At a typical mix of three input tokens to one output token, that is $0.205 per million tokens for Gemma 4 31B IT versus $0.263 for GPT OSS 120B (1.3× as much) and $0.299 for Qwen3 235B-A22B Instruct 2507 (1.5× as much).

Which scores higher on benchmarks?

Gemma 4 31B IT scores higher on the Capabilities Index (ECI): Gemma 4 31B IT 143.1 (#96 of 165), GPT OSS 120B 139.8 (#114 of 165) and Qwen3 235B-A22B Instruct 2507 138.8 (#118 of 165). The confidence ranges of the top two overlap (140.8–145.9 vs 135.6–142.2), so treat the gap as small.

Which is better for coding?

There are no published SWE-bench Verified results for Gemma 4 31B IT, GPT OSS 120B and Qwen3 235B-A22B Instruct 2507 yet, so there is no like-for-like coding score. On overall capability, Gemma 4 31B IT 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?

Gemma 4 31B IT and Qwen3 235B-A22B Instruct 2507 have the largest context windows (262,144 and 262,144 tokens), against 131,072 for GPT OSS 120B. Maximum output per response: Gemma 4 31B IT up to 32,768, GPT OSS 120B up to 32,768, Qwen3 235B-A22B Instruct 2507 up to 16,384 tokens.

Which can read images, PDFs, audio or video?

Gemma 4 31B IT accepts text and images; GPT OSS 120B accepts text; Qwen3 235B-A22B Instruct 2507 accepts text. Gemma 4 31B IT handles the widest range of inputs.

Are any of these open source?

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

Which is newer?

Gemma 4 31B IT is the newest, released Apr 2, 2026. GPT OSS 120B came out Aug 5, 2025; Qwen3 235B-A22B Instruct 2507 came out Jul 21, 2025. Knowledge cutoff: Gemma 4 31B IT Jan 2025, GPT OSS 120B 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