Comparing 2 models
Gemini 4 Carbon vs Kimi K3
Benchmarks through
Verdict
Gemini 4 Carbon and Kimi K3 are not yet ranked on the same task, so compare the specifications below. Kimi K3 costs 10.4x less per input token. Only Kimi K3 has open weights for self-hosting.
- Choose Kimi K3 for you need to self-host.
Built from the published task rankings below; ranks come from each task's own weighted benchmarks, so a win on one task says nothing about another.
Measured head to head
Benchmarks both models were run on
| Evaluation | Gemini 4 Carbon | Kimi K3 | Source |
|---|---|---|---|
| Design Arena Coding | 1348 Elo | 1352 Elo | source · 2026-11-15 |
| Design Arena UI | 1347 Elo | 1342 Elo | source · 2026-11-15 |
| Design Arena Website | 1319 Elo | 1336 Elo | source · 2026-11-15 |
| FrontierMath Tiers 1-3 (Epoch AI) | 90.5% ± 1.7 | 72.2% ± 2.7 | source · 2026-11-02 |
| GPQA Diamond (Epoch AI) | 96.5% ± 1.2 | 93.1% ± 1.5 | source · 2026-11-02 |
| LMArena Agent | 17 | 0.6 | source · 2026-11-11 |
| LMArena Agent Steerability | 19.1 | -0.8 | source · 2026-11-11 |
| LMArena Agent Task Outcome | 18.4 | 3.4 | source · 2026-11-11 |
| LMArena Agent Tool Hallucination | 0.5 | 1 | source · 2026-11-11 |
| LMArena Creative Writing | 1533 Elo | 1453 Elo | source · 2026-11-11 |
| LMArena Instruction Following | 1539 Elo | 1484 Elo | source · 2026-11-11 |
| LMArena Long Query | 1549 Elo | 1490 Elo | source · 2026-11-11 |
| LMArena Text | 1526 Elo | 1487 Elo | source · 2026-11-11 |
| LMArena Text Factuality | 1521 Elo | 1471 Elo | source · 2026-11-11 |
| LMArena WebDev | 1768 Elo | 1658 Elo | source · 2026-11-11 |
| OTIS Mock AIME (Epoch AI) | 100% ± 0 | 97.2% ± 1.1 | source · 2026-11-02 |
| SimpleQA Verified (Epoch AI) | 77.7% ± 1.3 | 50.6% ± 1.6 | source · 2026-11-02 |
| Terminal-Bench (4.1) | 64.5% | 23.0% | source · 2026-11-15 |
Values are as published by each evaluator, in its own units; rows from different evaluators are not on a common scale.
task evidence
How they place when the question changes
Lower ranks are better. “Best” marks the strongest option in this shortlist for that task.