ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Gemma 3 12B IT vs Qwen2.5 7B Instruct vs WizardLM 2 8x22B

Gemma 3 12B IT comes out ahead, 41 to 33 and 32 on our weighted score, and it is the cheaper option too.

Price in rating: Excluded · Include price

  1. Our pick

    Google

    Gemma 3 12B IT

    Released Mar 12, 2025

    41/100
    • ECI—
    • Price$0.05 / $0.15
    • Context131K
  2. Alibaba (Qwen)

    Qwen2.5 7B Instruct

    Released Sep 19, 2024

    33/100
    • ECI—
    • Price$0.075 / $0.15
    • Context131K
  3. Microsoft

    WizardLM 2 8x22B

    Released Apr 15, 2024

    32/100
    • ECI—
    • Price$0.493 / $0.493
    • Context66K
01 — Verdict

Gemma 3 12B IT is our pick

Gemma 3 12B IT is the better all-round choice, scoring 41/100 against Qwen2.5 7B Instruct (33) and WizardLM 2 8x22B (32). It leads on inputs & features. WizardLM 2 8x22B wins on capability. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond), because Gemma 3 12B IT, Qwen2.5 7B Instruct and WizardLM 2 8x22B has no Capabilities Index score yet.

  • CapabilityWizardLM 2 8x22BShared benchmarks: WizardLM 2 8x22B 43.4% · Gemma 3 12B IT 39.5% · Qwen2.5 7B Instruct 35.5%
  • Lowest priceGemma 3 12B ITGemma 3 12B IT $0.075 · Qwen2.5 7B Instruct $0.094 · WizardLM 2 8x22B $0.493 per 1M tokens (3:1 blend)
  • Longest contextGemma 3 12B IT and Qwen2.5 7B InstructGemma 3 12B IT 131,072 · Qwen2.5 7B Instruct 131,072 · WizardLM 2 8x22B 65,536 tokens
  • Widest inputsGemma 3 12B ITGemma 3 12B IT: Text, Images · Qwen2.5 7B Instruct: Text · WizardLM 2 8x22B: Text
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightGemma 3 12B ITQwen2.5 7B InstructWizardLM 2 8x22B
CapabilityShared benchmarks67%393543
Inputs & features20%50250
Context window13%323222
Overall100%41/10033/10032/100
02 — Side by side

Every spec in one table

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

Gemma 3 12B IT vs Qwen2.5 7B Instruct vs WizardLM 2 8x22B specifications side by side
SpecificationGemma 3 12B ITGoogleQwen2.5 7B InstructAlibaba (Qwen)WizardLM 2 8x22BMicrosoft
Capability
Capabilities Index (ECI)———
ECI rank———
GPQA DiamondGraduate-level science questions39.5%35.5%43.4% (best)
OTIS Mock AIME 2024–2025Competition mathematics16.7% (best)2.5%—
Price per million tokens
Input$0.05 (best)$0.075$0.493
Output$0.15 (best)$0.15 (best)$0.493
Cached input———
Blended (3:1)$0.075 (best)$0.094$0.493
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 13 providersMedian of 4 providersMedian of 3 providers
Limits
Context window131,072 tokens (best)131,072 tokens (best)65,536 tokens
Max output131,072 tokens (best)8,192 tokens8,000 tokens
Inputs and features
TextYesYesYes
ImagesYesNoNo
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningNoNoNo
Tool callingYesYesNo
Structured outputNoNoNo
Availability
WeightsOpenOpenApache 2.0OpenApache-2.0
API model ID———
API providers14 (best)53
ReleasedMar 12, 2025Sep 19, 2024Apr 15, 2024
Knowledge cutoffAug 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 3 12B IT$0.80
  • Qwen2.5 7B Instruct$1.05
  • WizardLM 2 8x22B$5.92
04 — Questions

Which should you choose?

Which is better: Gemma 3 12B IT, Qwen2.5 7B Instruct or WizardLM 2 8x22B?

Gemma 3 12B IT is the better all-round choice, scoring 41/100 against Qwen2.5 7B Instruct (33) and WizardLM 2 8x22B (32). It leads on inputs & features. WizardLM 2 8x22B wins on capability. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond), because Gemma 3 12B IT, Qwen2.5 7B Instruct and WizardLM 2 8x22B has no Capabilities Index score yet.

Which is cheaper, Gemma 3 12B IT, Qwen2.5 7B Instruct or WizardLM 2 8x22B?

Gemma 3 12B IT is cheaper at $0.05 input / $0.15 output per million tokens (median across 13 API providers). Qwen2.5 7B Instruct costs $0.075 input / $0.15 output per million tokens (median across 4 API providers); WizardLM 2 8x22B costs $0.493 input / $0.493 output per million tokens (median across 3 API providers). At a typical mix of three input tokens to one output token, that is $0.075 per million tokens for Gemma 3 12B IT versus $0.094 for Qwen2.5 7B Instruct (1.3× as much) and $0.493 for WizardLM 2 8x22B (6.6× as much).

Which scores higher on benchmarks?

Not every model here has a Capabilities Index score, so we compare the average of shared benchmarks (GPQA Diamond): WizardLM 2 8x22B 43.4%, Gemma 3 12B IT 39.5% and Qwen2.5 7B Instruct 35.5%. On individual benchmarks: GPQA Diamond — WizardLM 2 8x22B 43.4%, Gemma 3 12B IT 39.5%, Qwen2.5 7B Instruct 35.5%.

Which is better for coding?

There are no published SWE-bench Verified results for Gemma 3 12B IT, Qwen2.5 7B Instruct and WizardLM 2 8x22B yet, so there is no like-for-like coding score. On overall capability, WizardLM 2 8x22B leads, which tends to carry over to coding, but test on your own codebase. Note that WizardLM 2 8x22B does not support tool calling, which most coding agents need.

Which has the bigger context window?

Gemma 3 12B IT and Qwen2.5 7B Instruct have the largest context windows (131,072 and 131,072 tokens), against 65,536 for WizardLM 2 8x22B. Maximum output per response: Gemma 3 12B IT up to 131,072, Qwen2.5 7B Instruct up to 8,192, WizardLM 2 8x22B up to 8,000 tokens.

Which can read images, PDFs, audio or video?

Gemma 3 12B IT accepts text and images; Qwen2.5 7B Instruct accepts text; WizardLM 2 8x22B accepts text. Gemma 3 12B IT handles the widest range of inputs.

Are any of these open source?

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

Which is newer?

Gemma 3 12B IT is the newest, released Mar 12, 2025. Qwen2.5 7B Instruct came out Sep 19, 2024; WizardLM 2 8x22B came out Apr 15, 2024. Knowledge cutoff: Gemma 3 12B IT Aug 2024.

How do you decide the winner?

This rating uses capability (67%, independent benchmark results); inputs & features (20%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (13%, 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.