ZMIME
Comparison · 3 models · Updated Nov 18, 2026

Kimi K2.6 vs o3 vs Claude Sonnet 4.5

Kimi K2.6 comes out ahead, 74 to 71 and 71 on our weighted score, and it is the cheaper option too.

Price in rating: Excluded · Include price

  1. Our pick

    Moonshot AI

    Kimi K2.6

    Released Apr 20, 2026

    74/100
    • ECI151.6
    • Price$0.95 / $4.00
    • Context262K
  2. OpenAI

    o3

    Released Apr 16, 2025

    71/100
    • ECI147.1
    • Price$2.00 / $8.00
    • Context200K
  3. Anthropic

    Claude Sonnet 4.5

    Released Sep 29, 2025Deprecated

    71/100
    • ECI147.6
    • Price$3.00 / $15.00
    • Context200K
01 — Verdict

Kimi K2.6 is our pick

Kimi K2.6 is the better all-round choice, scoring 74/100 against Claude Sonnet 4.5 (71) and o3 (71). It leads on capability and context window. The score weighs capability 67%, inputs & features 20%, context window 13%.

  • CapabilityKimi K2.6Capabilities Index (ECI): Kimi K2.6 151.6 · Claude Sonnet 4.5 147.6 · o3 147.1
  • Lowest priceKimi K2.6Kimi K2.6 $1.71 · o3 $3.50 · Claude Sonnet 4.5 $6.00 per 1M tokens (3:1 blend)
  • Longest contextKimi K2.6Kimi K2.6 262,144 · o3 200,000 · Claude Sonnet 4.5 200,000 tokens
  • Widest inputsSame inputsKimi K2.6: Text, Images, Video · o3: Text, Images, PDFs · Claude Sonnet 4.5: Text, Images, PDFs
  • Self-hostingKimi K2.6Publishes downloadable weights
How the score is built
MeasureWeightKimi K2.6o3Claude Sonnet 4.5
CapabilityCapabilities Index (ECI)67%797575
Inputs & features20%808080
Context window13%433939
Overall100%74/10071/10071/100
02 — Side by side

Every spec in one table

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

Kimi K2.6 vs o3 vs Claude Sonnet 4.5 specifications side by side
SpecificationKimi K2.6Moonshot AIo3OpenAIClaude Sonnet 4.5Anthropic
Capability
Capabilities Index (ECI)151.6 (best)147.1147.6
ECI rank#61 of 158 (best)#81 of 158#78 of 158
GPQA DiamondGraduate-level science questions90.8% (best)81.8%82.3%
FrontierMath Tiers 1–3Research-level mathematics57.2% (best)33.3%23.9%
OTIS Mock AIME 2024–2025Competition mathematics96.1% (best)84.4%77.8%
SWE-bench VerifiedFixing real GitHub issues76.7% (best)62.3%71.3%
SimpleQA VerifiedShort factual questions34.9%49.4% (best)30.7%
Price per million tokens
Input$0.95 (best)$2.00$3.00
Output$4.00 (best)$8.00$15.00
Cached input$0.16 (best)$0.50$0.30
Blended (3:1)$1.71 (best)$3.50$6.00
Long-context rateSame rateSame rateSame rate
Price sourceOfficial Moonshot AI APIOfficial OpenAI APIOfficial Anthropic API
Limits
Context window262,144 tokens (best)200,000 tokens200,000 tokens
Max output262,144 tokens (best)100,000 tokens64,000 tokens
Inputs and features
TextYesYesYes
ImagesYesYesYes
PDFsNoYesYes
AudioNoNoNo
VideoYesNoNo
ReasoningYesYeslow · medium · highYes
Tool callingYesYesYes
Structured outputYesYesYes
Availability
WeightsOpenProprietaryProprietary
API model IDkimi-k2.6o3claude-sonnet-4-5
API providers64 (best)2340
ReleasedApr 20, 2026Apr 16, 2025Sep 29, 2025
Knowledge cutoffJan 2025May 2024Jul 31, 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.

  • Kimi K2.6$17.50
  • o3$36.00
  • Claude Sonnet 4.5$60.00
04 — Questions

Which should you choose?

Which is better: Kimi K2.6, o3 or Claude Sonnet 4.5?

Kimi K2.6 is the better all-round choice, scoring 74/100 against Claude Sonnet 4.5 (71) and o3 (71). It leads on capability and context window. The score weighs capability 67%, inputs & features 20%, context window 13%.

Which is cheaper, Kimi K2.6, o3 or Claude Sonnet 4.5?

Kimi K2.6 is cheaper at $0.95 input / $4.00 output per million tokens (official Moonshot AI API price). o3 costs $2.00 input / $8.00 output per million tokens (official OpenAI API price); Claude Sonnet 4.5 costs $3.00 input / $15.00 output per million tokens (official Anthropic 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 $3.50 for o3 (2× as much) and $6.00 for Claude Sonnet 4.5 (3.5× as much).

Which scores higher on benchmarks?

Kimi K2.6 scores higher on the Capabilities Index (ECI): Kimi K2.6 151.6 (#61 of 158), Claude Sonnet 4.5 147.6 (#78 of 158) and o3 147.1 (#81 of 158). Their confidence ranges do not overlap (149.9–153.4 vs 146.1–149.0), so the gap is a real one. On individual benchmarks: GPQA Diamond — Kimi K2.6 90.8%, Claude Sonnet 4.5 82.3%, o3 81.8%; FrontierMath Tiers 1–3 — Kimi K2.6 57.2%, o3 33.3%, Claude Sonnet 4.5 23.9%; OTIS Mock AIME 2024–2025 — Kimi K2.6 96.1%, o3 84.4%, Claude Sonnet 4.5 77.8%; SWE-bench Verified — Kimi K2.6 76.7%, Claude Sonnet 4.5 71.3%, o3 62.3%; SimpleQA Verified — o3 49.4%, Kimi K2.6 34.9%, Claude Sonnet 4.5 30.7%.

Which is better for coding?

Kimi K2.6 resolves more real GitHub issues on SWE-bench Verified: Kimi K2.6 76.7%, Claude Sonnet 4.5 71.3% and o3 62.3%. All three support tool calling for agent workflows.

Which has the bigger context window?

Kimi K2.6 has the largest context window at 262,144 tokens, against 200,000 for o3 and 200,000 for Claude Sonnet 4.5. Maximum output per response: Kimi K2.6 up to 262,144, o3 up to 100,000, Claude Sonnet 4.5 up to 64,000 tokens.

Which can read images, PDFs, audio or video?

Kimi K2.6 accepts text, images and video; o3 accepts text, images and PDFs; Claude Sonnet 4.5 accepts text, images and PDFs. They handle the same number of input types.

Are any of these open source?

Kimi K2.6 publishes its weights and can be self-hosted; o3 and Claude Sonnet 4.5 is proprietary.

Which is newer?

Kimi K2.6 is the newest, released Apr 20, 2026. Claude Sonnet 4.5 came out Sep 29, 2025; o3 came out Apr 16, 2025. Knowledge cutoff: Kimi K2.6 Jan 2025, o3 May 2024, Claude Sonnet 4.5 Jul 31, 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.