Kimi K3 · Data analysis
Kimi K3 for data analysis: our test runs
Kimi K3 ran our 5 data analysis prompts in llmwise, through the same pipeline your messages take. Here's every reply as it came, whether it passed, what it cost and what it counts as on Pro, next to the other models.
Based on 5 of our test runs on , through OpenRouter with the app's own prompt and settings. Updated .
Short answer
In our test runs on September 27, 2026, Kimi K3 passed 5 of 5 of our data analysis prompts (both hard ones): tied 1st of 20 models. Each reply counted as 1 message on Pro, where it has 125 a month, and cost $0.0068 to run on average.
Prompt by prompt
Each of the 5 data analysis prompts, and how Kimi K3's reply did.
| Prompt | Result | Cost | Time |
|---|---|---|---|
| The region with the most revenueEveryday | Passed: Final answer North, 1,085.00: right. | $0.0049 | 2.2 s |
| Average order value in AugustEveryday | Passed: Final answer 300.00: right. | $0.0054 | 2.7 s |
| Revenue change from July to AugustEveryday | Passed: Final answer 67.4%: right. | $0.0064 | 6.4 s |
| A median, filtered two waysHard | Passed: Final answer 6.25: right. | $0.0042 | 3.5 s |
| Correlation between ad spend and sign-upsHard | Passed: Final answer 0.97: right. | $0.0131 | 20.2 s |
Every data analysis reply from Kimi K3
The region with the most revenue
An everyday prompt: “The region with the most revenue”, in full.
Passed: Final answer North, 1,085.00: right.
**West:** 2×45 + 1×310 + 2×120 + 5×45 = 90 + 310 + 240 + 225 = **865** **East:** 1×120 + 3×45 + 1×310 + 3×120 = 120 + 135 + 310 + 360 = **925** **North:** 4×45 + 2×120 + 2×310 + 1×45 = 180 + 240 + 620 + 45 = **1085** Final answer: North, $1,085.00
728 tokens in, 183 out (45 of them reasoning) · 2.2 s · $0.0049 · 1 message on Pro · answered by moonshotai/kimi-k3 via Together ·
Average order value in August
An everyday prompt: “Average order value in August”, in full.
Passed: Final answer 300.00: right.
August 2026 orders: 1007–1012 Revenues: 240, 310, 620, 225, 360, 45 Total = 1800; orders = 6; average = 1800 / 6 = 300.00 Final answer: 300.00
735 tokens in, 234 out (155 of them reasoning) · 2.7 s · $0.0054 · 1 message on Pro · answered by moonshotai/kimi-k3 via Together ·
Revenue change from July to August
An everyday prompt: “Revenue change from July to August”, in full.
Passed: Final answer 67.4%: right.
July revenue: 90 + 120 + 310 + 180 + 135 + 240 = 1075 August revenue: 240 + 310 + 620 + 225 + 360 + 45 = 1800 Change: (1800 − 1075) / 1075 × 100 = 67.441...% ≈ 67.4% increase Final answer: 67.4%
740 tokens in, 415 out (314 of them reasoning) · 6.4 s · $0.0064 · 1 message on Pro · answered by moonshotai/kimi-k3 via Phala ·
A median, filtered two ways
A hard prompt: “A median, filtered two ways”, in full.
Passed: Final answer 6.25: right.
Billing + high tickets: 5.5, 2, 9, 7, 26, 4 Sorted: 2, 4, 5.5, 7, 9, 26 Median = (5.5 + 7) / 2 = 6.25 Final answer: 6.25
597 tokens in, 265 out (182 of them reasoning) · 3.5 s · $0.0042 · 1 message on Pro · answered by moonshotai/kimi-k3 via Phala ·
Correlation between ad spend and sign-ups
A hard prompt: “Correlation between ad spend and sign-ups”, in full.
Passed: Final answer 0.97: right.
## Working **Means:** x̄ = 6150/8 = 768.75, ȳ = 434/8 = 54.25 **Deviations and sums:** | week | x−x̄ | y−ȳ | (x−x̄)(y−ȳ) | (x−x̄)² | (y−ȳ)² | |---|---|---|---|---|---| | 1 | −268.75 | −12.25 | 3292.19 | 72226.56 | 150.06 | | 2 | 31.25 | 0.75 | 23.44 | 976.56 | 0.56 | | 3 | −118.75 | −5.25 | 623.44 | 14101.56 | 27.56 | | 4 | 431.25 | 16.75 | 7223.44 | 185976.56 | 280.56 | | 5 | 131.25 | 5.75 | 754.69 | 17226.56 | 33.06 | | 6 | −468.75 | −19.25 | 9023.44 | 219726.56 | 370.56 | | 7 | 331.25 | 9.75 | 3229.69 | 109726.56 | 95.06 | | 8 | −68.75 | 3.75 | −257.81 | 4726.56 | 14.06 | | **Σ** | | | **23912.50** | **624687.50** | **971.50** | **Compute r:** r = 23912.5 / √(624687.5 × 971.5) = 23912.5 / √(606,883,906.25) = 23912.5 / 24635.0 ≈ 0.9707 Rounded to 2 decimal places: 0.97 (a very strong positive correlation). Final answer: 0.97
537 tokens in, 1,418 out (964 of them reasoning) · 20.2 s · $0.0131 · 1 message on Pro · answered by moonshotai/kimi-k3 via Wafer ·
How it compares on data analysis
Kimi K3 beside the job's picks among every model. 16 models share the hard-problems pick, Kimi K3 among them.
| Model | Passed | On Pro | Cost per reply |
|---|---|---|---|
| GLM 5.3Our pick: hard problems (shared), best value | 5 of 5 | 250 a month on Pro | $0.0019 |
| GPT-6 LunaOur pick: hard problems (shared), everyday | 5 of 5 | 60 a day on Pro | $0.0002 |
| Kimi K3Our pick: hard problems (shared) | 5 of 5 | 125 a month on Pro | $0.0068 |
Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Plus $3.75, Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.
Kimi K3 next to each model on data analysis
Kimi K3 against each model on the same 5 data analysis prompts. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.
Against Claude Fable 5.1: Kimi K3 answered 1.4× sooner (3.5 s to 4.9 s), cost 4.2× less, and passed 5 of 5 to its 5.
Against Claude Opus 5.5: Kimi K3 answered 1.3× sooner (3.5 s to 4.6 s), cost 1.9× less, and passed 5 of 5 to its 5.
Against Claude Sonnet 5.5: Kimi K3 answered 1.5× later (3.5 s to 2.3 s), cost 1.4× more, and passed 5 of 5 to its 5.
Against Claude Sonnet 5: Kimi K3 answered 1.5× sooner (3.5 s to 5.3 s), cost within 10% of it, and passed 5 of 5 to its 4. Only Kimi K3 passed Correlation between ad spend and sign-ups.
Against Claude Haiku 5.5: Kimi K3 answered within 10% of its time (3.5 s to 3.3 s), cost 18.9× more, and passed 5 of 5 to its 5.
Against Claude Haiku 4.5: Kimi K3 answered 1.3× later (3.5 s to 2.7 s), cost 2.7× more, and passed 5 of 5 to its 3. Only Kimi K3 passed Average order value in August; Correlation between ad spend and sign-ups.
Against GPT-6 Astra: Kimi K3 answered within 10% of its time (3.5 s to 3.3 s), cost 2.3× less, and passed 5 of 5 to its 5.
Against GPT-6.1 Sol: Kimi K3 answered 1.2× later (3.5 s to 3.0 s), cost 4.3× more, and passed 5 of 5 to its 5.
Against GPT-6 Sol: Kimi K3 answered 1.3× later (3.5 s to 2.8 s), cost 2.0× more, and passed 5 of 5 to its 5.
Against GPT-6 Luna: Kimi K3 answered 1.4× later (3.5 s to 2.5 s), cost 38.5× more, and passed 5 of 5 to its 5.
Against Gemini 4 Argon: Kimi K3 answered 2.0× sooner (3.5 s to 6.9 s), cost 2.4× more, and passed 5 of 5 to its 5.
Against Gemini 3.1 Pro (preview): Kimi K3 answered 2.5× sooner (3.5 s to 8.6 s), cost 2.3× less, and passed 5 of 5 to its 5.
Against Gemini 3.8 Flash: Kimi K3 answered 1.5× sooner (3.5 s to 5.4 s), cost 2.5× more, and passed 5 of 5 to its 4. Only Kimi K3 passed Correlation between ad spend and sign-ups.
Against DeepSeek V4.1 Flash: Kimi K3 answered 4.1× later (3.5 s to 0.9 s), cost 8.0× more, and passed 5 of 5 to its 5.
Against DeepSeek V4 Pro: Kimi K3 answered 1.2× sooner (3.5 s to 4.1 s), cost 2.4× more, and passed 5 of 5 to its 5.
Against Grok 4.7: Kimi K3 answered 2.5× sooner (3.5 s to 8.9 s), cost 1.6× less, and passed 5 of 5 to its 5.
Against GLM 5.3: Kimi K3 answered 1.8× later (3.5 s to 1.9 s), cost 3.5× more, and passed 5 of 5 to its 5.
Against GLM 5.3 Flash: Kimi K3 answered 2.8× sooner (3.5 s to 9.8 s), cost 17.6× more, and passed 5 of 5 to its 4. Only Kimi K3 passed Correlation between ad spend and sign-ups.
Against Mistral Large 4: Kimi K3 answered 4.5× sooner (3.5 s to 15.8 s), cost 1.8× more, and passed 5 of 5 to its 5.
How these runs were done
Final answer. Automatic. The reply's last “Final answer:” line must hold the right value.
Kimi K3, and data analysis, elsewhere
- Kimi K3: price, limits and messages on every plan
- Kimi K3 for coding: our test runs
- Kimi K3 for writing: our test runs
- Kimi K3 for math: our test runs
- Claude Sonnet 5.5 vs Kimi K3
- Gemini 4 Argon vs Kimi K3
- Claude Fable 5.1 vs Kimi K3
- Kimi K3 vs GLM 5.3
- DeepSeek V4 Pro vs Kimi K3
- Claude Opus 5.5 vs Kimi K3
- Claude Sonnet 5 vs Kimi K3
- Claude Haiku 4.5 vs Kimi K3
- GPT-6 Sol vs Kimi K3
- Gemini 3.1 Pro (preview) vs Kimi K3
- Gemini 3.8 Flash vs Kimi K3
- Grok 4.7 vs Kimi K3
- Analyze a CSV with code: the prompt, and its price
- AI prompts for business, tested on four models
- Our test runs: 50 prompts on every model
Questions
Is Kimi K3 good for data analysis?
In our test runs it passed 5 of 5 data analysis prompts, tied 1st of the 20 models in llmwise. Every reply is on this page, so you can judge them yourself.
How many of my messages does a data analysis reply from Kimi K3 use?
1 message each on Pro, where it has 125 a month on Pro. The price of a message is fixed and shown before you send it, however long the reply.
How were these runs done?
The same way for every model: each prompt sent through llmwise's own pipeline, each reply checked the same way. The methods page has every prompt and how each is scored.
Claude, GPT, Gemini, DeepSeek, Grok, Kimi, GLM, and Mistral, in one chat.
See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.