ZMIME
Comparison · 3 models · Updated Nov 18, 2026

DeepSeek V4 Pro vs Kimi K2.5 vs Qwen3.6 Max Preview

Kimi K2.5 comes out ahead, 70 to 67 and 64 on our weighted score, and it is the cheaper option too.

Price in rating: Excluded · Include price

  1. DeepSeek

    DeepSeek V4 Pro

    Released Apr 24, 2026

    67/100
    • ECI150.2
    • Price$1.60 / $3.33
    • Context1M
  2. Our pick

    Moonshot AI

    Kimi K2.5

    Released Jan 27, 2026

    70/100
    • ECI148.8
    • Price$0.60 / $3.00
    • Context262K
  3. Alibaba (Qwen)

    Qwen3.6 Max Preview

    Released Apr 20, 2026Deprecated

    64/100
    • ECI149.9
    • Price$1.30 / $7.80
    • Context262K
01 — Verdict

Kimi K2.5 is our pick

Kimi K2.5 is the better all-round choice, scoring 70/100 against DeepSeek V4 Pro (67) and Qwen3.6 Max Preview (64). It leads on inputs & features. DeepSeek V4 Pro wins on context window. The score weighs capability 67%, inputs & features 20%, context window 13%.

  • CapabilityDeepSeek V4 ProCapabilities Index (ECI): DeepSeek V4 Pro 150.2 · Qwen3.6 Max Preview 149.9 · Kimi K2.5 148.8
  • Lowest priceKimi K2.5Kimi K2.5 $1.20 · DeepSeek V4 Pro $2.03 · Qwen3.6 Max Preview $2.92 per 1M tokens (3:1 blend)
  • Longest contextDeepSeek V4 ProDeepSeek V4 Pro 1,000,000 · Kimi K2.5 262,144 · Qwen3.6 Max Preview 262,144 tokens
  • Widest inputsKimi K2.5DeepSeek V4 Pro: Text · Kimi K2.5: Text, Images, Video · Qwen3.6 Max Preview: Text
  • Self-hostingDeepSeek V4 Pro and Kimi K2.5Publishes downloadable weights
How the score is built
MeasureWeightDeepSeek V4 ProKimi K2.5Qwen3.6 Max Preview
CapabilityCapabilities Index (ECI)67%777677
Inputs & features20%357035
Context window13%644343
Overall100%67/10070/10064/100
02 — Side by side

Every spec in one table

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

DeepSeek V4 Pro vs Kimi K2.5 vs Qwen3.6 Max Preview specifications side by side
SpecificationDeepSeek V4 ProDeepSeekKimi K2.5Moonshot AIQwen3.6 Max PreviewAlibaba (Qwen)
Capability
Capabilities Index (ECI)150.2 (best)148.8149.9
ECI rank#67 of 158 (best)#73 of 158#70 of 158
GPQA DiamondGraduate-level science questions90.9% (best)87.6%87.4%
FrontierMath Tiers 1–3Research-level mathematics45.3%——
OTIS Mock AIME 2024–2025Competition mathematics96.7% (best)92.2%91.1%
SWE-bench VerifiedFixing real GitHub issues77.6% (best)73.8%76.7%
SimpleQA VerifiedShort factual questions47.0%34.3%52.0% (best)
Price per million tokens
Input$1.60$0.60 (best)$1.30
Output$3.33$3.00 (best)$7.80
Cached input——$0.13
Blended (3:1)$2.03$1.20 (best)$2.92
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 65 providersMedian of 36 providersOfficial Alibaba API
Limits
Context window1,000,000 tokens (best)262,144 tokens262,144 tokens
Max output384,000 tokens (best)262,144 tokens65,536 tokens
Inputs and features
TextYesYesYes
ImagesNoYesNo
PDFsNoNoNo
AudioNoNoNo
VideoNoYesNo
ReasoningYesYesYes
Tool callingYesYesYes
Structured outputNoNoNo
Availability
WeightsOpenOpenProprietary
API model ID——qwen3.6-max-preview
API providers67 (best)365
ReleasedApr 24, 2026Jan 27, 2026Apr 20, 2026
Knowledge cutoffMay 2025Jan 2025Apr 2025
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.

  • DeepSeek V4 Pro$22.66
  • Kimi K2.5$12.00
  • Qwen3.6 Max Preview$28.60
04 — Questions

Which should you choose?

Which is better: DeepSeek V4 Pro, Kimi K2.5 or Qwen3.6 Max Preview?

Kimi K2.5 is the better all-round choice, scoring 70/100 against DeepSeek V4 Pro (67) and Qwen3.6 Max Preview (64). It leads on inputs & features. DeepSeek V4 Pro wins on context window. The score weighs capability 67%, inputs & features 20%, context window 13%.

Which is cheaper, DeepSeek V4 Pro, Kimi K2.5 or Qwen3.6 Max Preview?

Kimi K2.5 is cheaper at $0.60 input / $3.00 output per million tokens (median across 36 API providers). DeepSeek V4 Pro costs $1.60 input / $3.33 output per million tokens (median across 65 API providers); Qwen3.6 Max Preview costs $1.30 input / $7.80 output per million tokens (official Alibaba API price). At a typical mix of three input tokens to one output token, that is $1.20 per million tokens for Kimi K2.5 versus $2.03 for DeepSeek V4 Pro (1.7× as much) and $2.92 for Qwen3.6 Max Preview (2.4× as much).

Which scores higher on benchmarks?

DeepSeek V4 Pro scores higher on the Capabilities Index (ECI): DeepSeek V4 Pro 150.2 (#67 of 158), Qwen3.6 Max Preview 149.9 (#70 of 158) and Kimi K2.5 148.8 (#73 of 158). The confidence ranges of the top two overlap (148.5–152.0 vs 147.9–152.1), so treat the gap as small. On individual benchmarks: GPQA Diamond — DeepSeek V4 Pro 90.9%, Kimi K2.5 87.6%, Qwen3.6 Max Preview 87.4%; OTIS Mock AIME 2024–2025 — DeepSeek V4 Pro 96.7%, Kimi K2.5 92.2%, Qwen3.6 Max Preview 91.1%; SWE-bench Verified — DeepSeek V4 Pro 77.6%, Qwen3.6 Max Preview 76.7%, Kimi K2.5 73.8%; SimpleQA Verified — Qwen3.6 Max Preview 52.0%, DeepSeek V4 Pro 47.0%, Kimi K2.5 34.3%.

Which is better for coding?

DeepSeek V4 Pro resolves more real GitHub issues on SWE-bench Verified: DeepSeek V4 Pro 77.6%, Qwen3.6 Max Preview 76.7% and Kimi K2.5 73.8%. All three support tool calling for agent workflows.

Which has the bigger context window?

DeepSeek V4 Pro has the largest context window at 1,000,000 tokens, against 262,144 for Kimi K2.5 and 262,144 for Qwen3.6 Max Preview. Maximum output per response: DeepSeek V4 Pro up to 384,000, Kimi K2.5 up to 262,144, Qwen3.6 Max Preview up to 65,536 tokens.

Which can read images, PDFs, audio or video?

DeepSeek V4 Pro accepts text; Kimi K2.5 accepts text, images and video; Qwen3.6 Max Preview accepts text. Kimi K2.5 handles the widest range of inputs.

Are any of these open source?

DeepSeek V4 Pro and Kimi K2.5 publishes its weights and can be self-hosted; Qwen3.6 Max Preview is proprietary.

Which is newer?

DeepSeek V4 Pro is the newest, released Apr 24, 2026. Qwen3.6 Max Preview came out Apr 20, 2026; Kimi K2.5 came out Jan 27, 2026. Knowledge cutoff: DeepSeek V4 Pro May 2025, Kimi K2.5 Jan 2025, Qwen3.6 Max Preview Apr 2025.

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.