DeepSeek V4.1 Flash vs Kimi K3
| Spec | DeepSeek V4.1 Flash | Kimi K3 |
|---|---|---|
| Respan Index | 52.5 (#36) | 52.9 (#34) |
| Blended price per 1M | $0.089 | $2.52 |
| Context window | 1M | 1.0M |
| Input / output per 1M | $0.051 / $0.203 | $0.192 / $9.5 |
| Max output | 384K | 1.0M |
| Input | text, vision | text, vision |
| Released | 2026-09-10 | 2026-07-16 |
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | DeepSeek V4.1 Flash | Kimi K3 |
|---|---|---|
| Coding | 43.7 DeepSeek V4.1 Flash | 38.8 Kimi K3 |
| Agents | 58.3 DeepSeek V4.1 Flash | 61.8 Kimi K3 |
| Reasoning | 55.5 DeepSeek V4.1 Flash | 54.8 Kimi K3 |
| Knowledge & language | 44.5 DeepSeek V4.1 Flash | 52.9 Kimi K3 |
| Human preference | 58.0 DeepSeek V4.1 Flash | 58.1 Kimi K3 |
Benchmarks both report
They split the 25 benchmarks both report evenly. Higher is better. Each score links to where it was published.
| Benchmark | DeepSeek V4.1 Flash | Kimi K3 |
|---|---|---|
| LMArena Agent (LMArena) | 0.0078 | -0.0015 |
| LMArena Elo | 1475 | 1487 |
| LMArena WebDev (LMArena) | 1659 | 1658 |
| MirrorCode (Epoch AI) | 64.4% | 54.4% |
| APEX-Agents (Mercor) | 39.5% | 47.8% |
| FrontierMath Tier 4 v2 (Epoch AI) | 26.8% | 39% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 67.4% | 72.2% |
| GPQA Diamond | 90.9% | 93.5% |
| Mystery Game Puzzles (Epoch AI) | 43% | 26% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 98.3% | 97.2% |
| Terminal-Bench 4.0 (Vals AI) | 19.7% | 17.2% |
| SAGE 1.1 (Vals AI) | 85.4% | 86.8% |
| Chess Puzzles (Epoch AI) | 48% | 37% |
| SimpleQA Verified | 42.9% | 50.6% |
| Furniture Assembly (Epoch AI) | 34.2% | 34.2% |
| LiveBench Agentic Coding (LiveBench) | 77.3% | 62.2% |
| LiveBench Coding (LiveBench) | 80% | 81.4% |
| LiveBench Data Analysis (LiveBench) | 79.3% | 78.7% |
| LiveBench Instruction Following (LiveBench) | 70% | 71.4% |
| LiveBench Language (LiveBench) | 81.2% | 85.5% |
| LiveBench Mathematics (LiveBench) | 93.3% | 84.4% |
| LiveBench Reasoning (LiveBench) | 86.7% | 90.7% |
| SAGE (Vals AI) | 47.9% | 52.8% |
| SimpleBench (SimpleBench) | 66.7% | 60.7% |
| Terminal-Bench 2.1 | 90.6% | 88.3% |
FAQ
Is DeepSeek V4.1 Flash better than Kimi K3?
Kimi K3 scores 52.9 on the Respan Index (#34), ahead of DeepSeek V4.1 Flash at 52.5 (#36). They split the 25 benchmarks both report evenly.
Which is cheaper, DeepSeek V4.1 Flash or Kimi K3?
DeepSeek V4.1 Flash lists at $0.051 input and $0.203 output per 1M tokens; Kimi K3 at $0.192 and $9.5. DeepSeek V4.1 Flash is 28x cheaper per token at list price.
Which has the larger context window?
DeepSeek V4.1 Flash accepts 1M tokens and Kimi K3 1.0M tokens.
Compare other models
- vsDeepSeek V4.1 Flash vs DeepSeek V4 Flash
- vsKimi K3 vs Kimi K2.6
- vsGemini 3.8 Flash vs DeepSeek V4.1 Flash
- vsGPT-6.1 Sol vs Kimi K3
- vsDeepSeek V4.1 Flash vs Qwen3.8-Max
- vsDeepSeek V4.1 Flash vs MiniMax M3.1 Flash
- vsDeepSeek V4.1 Flash vs GLM-5.3
- vsKimi K3 vs Grok 4.7
- vsKimi K3 vs Qwen3.8-Max
- vsKimi K3 vs Inkling 2 Small