ZMIME
Comparison · 3 models · Updated Nov 18, 2026

GLM-5.3 vs Kimi K2.6 vs Muse Spark Open

Muse Spark Open comes out ahead, 68 to 65 and 63 on our weighted score, though Kimi K2.6 is 14% cheaper per token.

Price in rating: Included · Exclude price

  1. Z.ai (Zhipu)

    GLM-5.3

    Released Aug 14, 2026

    63/100
    • ECI157.1
    • Price$1.40 / $4.40
    • Context1M
  2. Moonshot AI

    Kimi K2.6

    Released Apr 20, 2026

    65/100
    • ECI151.6
    • Price$0.95 / $4.00
    • Context262K
  3. Our pick

    Meta

    Muse Spark Open

    Released Oct 15, 2026New

    68/100
    • ECI155.5
    • Price$1.25 / $4.25
    • Context1.05M
01 — Verdict

Muse Spark Open is our pick

Muse Spark Open is the better all-round choice, scoring 68/100 against Kimi K2.6 (65) and GLM-5.3 (63). Kimi K2.6 wins on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

  • CapabilityGLM-5.3Capabilities Index (ECI): GLM-5.3 157.1 · Muse Spark Open 155.5 · Kimi K2.6 151.6
  • Lowest priceKimi K2.6Kimi K2.6 $1.71 · Muse Spark Open $2.00 · GLM-5.3 $2.15 per 1M tokens (3:1 blend)
  • Longest contextMuse Spark OpenMuse Spark Open 1,048,576 · GLM-5.3 1,000,000 · Kimi K2.6 262,144 tokens
  • Widest inputsMuse Spark OpenGLM-5.3: Text · Kimi K2.6: Text, Images, Video · Muse Spark Open: Text, Images, Audio, Video
  • Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
How the score is built
MeasureWeightGLM-5.3Kimi K2.6Muse Spark Open
CapabilityCapabilities Index (ECI)50%847982
Price25%333734
Inputs & features15%458080
Context window10%644365
Overall100%63/10065/10068/100
02 — Side by side

Every spec in one table

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

GLM-5.3 vs Kimi K2.6 vs Muse Spark Open specifications side by side
SpecificationGLM-5.3Z.ai (Zhipu)Kimi K2.6Moonshot AIMuse Spark OpenMeta
Capability
Capabilities Index (ECI)157.1 (best)151.6155.5
ECI rank#33 of 158 (best)#61 of 158#43 of 158
GPQA DiamondGraduate-level science questions90.9% (best)90.8%88.6%
FrontierMath Tiers 1–3Research-level mathematics68.8%57.2%69.8% (best)
OTIS Mock AIME 2024–2025Competition mathematics91.1%96.1%98.1% (best)
SWE-bench VerifiedFixing real GitHub issues—76.7%—
SimpleQA VerifiedShort factual questions41.0%34.9%56.6% (best)
Price per million tokens
Input$1.40$0.95 (best)$1.25
Output$4.40$4.00 (best)$4.25
Cached input$0.26$0.16 (best)—
Blended (3:1)$2.15$1.71 (best)$2.00
Long-context rateSame rateSame rateSame rate
Price sourceOfficial Z.AI APIOfficial Moonshot AI APIMedian of 22 providers
Limits
Context window1,000,000 tokens262,144 tokens1,048,576 tokens (best)
Max output131,072 tokens262,144 tokens (best)131,072 tokens
Inputs and features
TextYesYesYes
ImagesNoYesYes
PDFsNoNoNo
AudioNoNoYes
VideoNoYesYes
ReasoningYeslow · high · maxYesYes
Tool callingYesYesYes
Structured outputYesYesNo
Availability
WeightsOpenOpenOpenMuse Model License
API model IDglm-5.3kimi-k2.6—
API providers123 (best)6422
ReleasedAug 14, 2026Apr 20, 2026Oct 15, 2026
Knowledge cutoff—Jan 2025Jan 4, 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.

  • GLM-5.3$22.80
  • Kimi K2.6$17.50
  • Muse Spark Open$21.00
04 — Questions

Which should you choose?

Which is better: GLM-5.3, Kimi K2.6 or Muse Spark Open?

Muse Spark Open is the better all-round choice, scoring 68/100 against Kimi K2.6 (65) and GLM-5.3 (63). Kimi K2.6 wins on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.

Which is cheaper, GLM-5.3, Kimi K2.6 or Muse Spark Open?

Kimi K2.6 is cheaper at $0.95 input / $4.00 output per million tokens (official Moonshot AI API price). Muse Spark Open costs $1.25 input / $4.25 output per million tokens (median across 22 API providers); GLM-5.3 costs $1.40 input / $4.40 output per million tokens (official Z.AI API price). At a typical mix of three input tokens to one output token, that is $1.71 per million tokens for Kimi K2.6 versus $2.00 for Muse Spark Open (1.2× as much) and $2.15 for GLM-5.3 (1.3× as much).

Which scores higher on benchmarks?

GLM-5.3 scores higher on the Capabilities Index (ECI): GLM-5.3 157.1 (#33 of 158), Muse Spark Open 155.5 (#43 of 158) and Kimi K2.6 151.6 (#61 of 158). The confidence ranges of the top two overlap (155.0–159.3 vs 153.4–157.7), so treat the gap as small. On individual benchmarks: GPQA Diamond — GLM-5.3 90.9%, Kimi K2.6 90.8%, Muse Spark Open 88.6%; FrontierMath Tiers 1–3 — Muse Spark Open 69.8%, GLM-5.3 68.8%, Kimi K2.6 57.2%; OTIS Mock AIME 2024–2025 — Muse Spark Open 98.1%, Kimi K2.6 96.1%, GLM-5.3 91.1%; SimpleQA Verified — Muse Spark Open 56.6%, GLM-5.3 41.0%, Kimi K2.6 34.9%.

Which is better for coding?

There are no published SWE-bench Verified results for GLM-5.3 and Muse Spark Open yet, so there is no like-for-like coding score. On overall capability, GLM-5.3 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?

Muse Spark Open has the largest context window at 1,048,576 tokens, against 1,000,000 for GLM-5.3 and 262,144 for Kimi K2.6. Maximum output per response: GLM-5.3 up to 131,072, Kimi K2.6 up to 262,144, Muse Spark Open up to 131,072 tokens.

Which can read images, PDFs, audio or video?

GLM-5.3 accepts text; Kimi K2.6 accepts text, images and video; Muse Spark Open accepts text, images, audio and video. Muse Spark Open handles the widest range of inputs.

Are any of these open source?

Yes, all three publish their weights (Muse Model License), so you can self-host them.

Which is newer?

Muse Spark Open is the newest, released Oct 15, 2026. GLM-5.3 came out Aug 14, 2026; Kimi K2.6 came out Apr 20, 2026. Knowledge cutoff: Kimi K2.6 Jan 2025, Muse Spark Open Jan 4, 2026.

How do you decide the winner?

This rating uses 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 18, 2026.