ZMIME
Comparison · 2 models · Updated Nov 19, 2026

Gemini 4 Hydrogen vs Kimi K3.1 Pro

Gemini 4 Hydrogen comes out ahead, 88 to 80 on our weighted score, and it is the cheaper option too.

Price in rating: Excluded · Include price

  1. Our pick

    Google

    Gemini 4 Hydrogen

    Released Nov 12, 2026New

    88/100
    • ECI—
    • Price$1.50 / $7.50
    • Context1.05M
  2. Moonshot AI

    Kimi K3.1 Pro

    Released Oct 23, 2026New

    80/100
    • ECI161.4
    • Price$2.00 / $10.00
    • Context1.05M
  3. Add a model

    Make it a three-way comparison.

01 — Verdict

Gemini 4 Hydrogen is our pick

Gemini 4 Hydrogen is the better all-round choice, scoring 88/100 against Kimi K3.1 Pro (80). It leads on capability and inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond, FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified), because Gemini 4 Hydrogen has no Capabilities Index score yet.

  • CapabilityGemini 4 HydrogenShared benchmarks: Gemini 4 Hydrogen 88.9% · Kimi K3.1 Pro 83.7%
  • Lowest priceGemini 4 HydrogenGemini 4 Hydrogen $3.00 · Kimi K3.1 Pro $4.00 per 1M tokens (3:1 blend)
  • Longest contextAbout the sameGemini 4 Hydrogen 1,048,576 · Kimi K3.1 Pro 1,048,576 tokens
  • Widest inputsGemini 4 HydrogenGemini 4 Hydrogen: Text, Images, PDFs, Audio, Video · Kimi K3.1 Pro: Text, Images, Video
  • Self-hostingKimi K3.1 ProPublishes downloadable weights
How the score is built
MeasureWeightGemini 4 HydrogenKimi K3.1 Pro
CapabilityShared benchmarks67%8984
Inputs & features20%10080
Context window13%6565
Overall100%88/10080/100
02 — Side by side

Every spec in one table

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

Gemini 4 Hydrogen vs Kimi K3.1 Pro specifications side by side
SpecificationGemini 4 HydrogenGoogleKimi K3.1 ProMoonshot AI
Capability
Capabilities Index (ECI)—161.4
ECI rank—#17 of 157
GPQA DiamondGraduate-level science questions96.8% (best)94.2%
FrontierMath Tiers 1–3Research-level mathematics88.1% (best)78.3%
OTIS Mock AIME 2024–2025Competition mathematics100%100%
SimpleQA VerifiedShort factual questions70.9% (best)62.2%
Price per million tokens
Input$1.50 (best)$2.00
Output$7.50 (best)$10.00
Cached input$0.075 (best)$0.20
Blended (3:1)$3.00 (best)$4.00
Long-context rateSame rateSame rate
Price sourceOfficial Google APIOfficial Moonshot AI API
Limits
Context window1,048,576 tokens1,048,576 tokens
Max output65,536 tokens1,048,576 tokens (best)
Inputs and features
TextYesYes
ImagesYesYes
PDFsYesNo
AudioYesNo
VideoYesYes
ReasoningYeslow · medium · high · xhighYeslow · high · max
Tool callingYesYes
Structured outputYesYes
Availability
WeightsProprietaryOpen
API model IDgemini-4-hydrogenkimi-k3.1-pro
API providers1757 (best)
ReleasedNov 12, 2026Oct 23, 2026
Knowledge cutoffMar 2026—
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.

  • Gemini 4 Hydrogen$30.00
  • Kimi K3.1 Pro$40.00
04 — Questions

Which should you choose?

Which is better: Gemini 4 Hydrogen or Kimi K3.1 Pro?

Gemini 4 Hydrogen is the better all-round choice, scoring 88/100 against Kimi K3.1 Pro (80). It leads on capability and inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond, FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified), because Gemini 4 Hydrogen has no Capabilities Index score yet.

Which is cheaper, Gemini 4 Hydrogen or Kimi K3.1 Pro?

Gemini 4 Hydrogen is cheaper at $1.50 input / $7.50 output per million tokens (official Google API price). Kimi K3.1 Pro costs $2.00 input / $10.00 output per million tokens (official Moonshot AI API price). At a typical mix of three input tokens to one output token, that is $3.00 per million tokens for Gemini 4 Hydrogen versus $4.00 for Kimi K3.1 Pro (1.3× 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, FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified): Gemini 4 Hydrogen 88.9% and Kimi K3.1 Pro 83.7%. On individual benchmarks: GPQA Diamond — Gemini 4 Hydrogen 96.8%, Kimi K3.1 Pro 94.2%; FrontierMath Tiers 1–3 — Gemini 4 Hydrogen 88.1%, Kimi K3.1 Pro 78.3%; OTIS Mock AIME 2024–2025 — Gemini 4 Hydrogen 100%, Kimi K3.1 Pro 100%; SimpleQA Verified — Gemini 4 Hydrogen 70.9%, Kimi K3.1 Pro 62.2%.

Which is better for coding?

There are no published SWE-bench Verified results for Gemini 4 Hydrogen and Kimi K3.1 Pro yet, so there is no like-for-like coding score. On overall capability, Gemini 4 Hydrogen leads, which tends to carry over to coding, but test on your own codebase. Both support tool calling for agent workflows.

Which has the bigger context window?

Gemini 4 Hydrogen and Kimi K3.1 Pro share the same 1,048,576-token context window. Maximum output per response: Gemini 4 Hydrogen up to 65,536, Kimi K3.1 Pro up to 1,048,576 tokens.

Which can read images, PDFs, audio or video?

Gemini 4 Hydrogen accepts text, images, PDFs, audio and video; Kimi K3.1 Pro accepts text, images and video. Gemini 4 Hydrogen handles the widest range of inputs.

Are any of these open source?

Kimi K3.1 Pro publishes its weights and can be self-hosted; Gemini 4 Hydrogen is proprietary.

Which is newer?

Gemini 4 Hydrogen is the newest, released Nov 12, 2026. Kimi K3.1 Pro came out Oct 23, 2026. Knowledge cutoff: Gemini 4 Hydrogen Mar 2026.

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 19, 2026.