Compare / MiniMax M2.1 vs Kimi K2 Thinking
MiniMax M2.1 vs Kimi K2 Thinking
Which is cheaper, which is more capable, and which does better at each kind of work, from 16 results on 3 sources that measured both models.
MiniMax M2.1 or Kimi K2 Thinking? The short answer
- MiniMax M2.1 is cheaper: $0.52 against $1.07 per million tokens, 2.0 times less.
- Kimi K2 Thinking scores higher on the Artificial Analysis Intelligence Index: 22.0 against 20.9.
- On our business benchmarks, MiniMax M2.1 ranks higher for agents and tool use.
- On our business benchmarks, Kimi K2 Thinking ranks higher for reasoning and knowledge and writing.
At a glance
| MiniMax M2.1 | Kimi K2 Thinking | |
|---|---|---|
| Price per million tokens | $0.30 in · $1.20 out | $0.60 in · $2.50 out |
| Blended (3 in : 1 out) | $0.52 | $1.07 |
| Cheapest host, blended | $0.52 (Novita, FP8) | $1.07 (Google) |
| Intelligence Index | 20.9 | 22.0 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | – | – |
| Context window | 204,800 tokens | 262,144 tokens |
| Weights | Open | Open |
| Developer based in | China | China |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among models available today, for each kind of work, from our own benchmarks and licensed sources. The better rank is in green.
| Use case | MiniMax M2.1 | Kimi K2 Thinking |
|---|---|---|
| Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls | 125 of 161 | 133 of 161 |
| Reasoning and knowledgeHard questions across science, maths and general knowledge | 86 of 195 | 82 of 195 |
| WritingWhat people prefer in blind comparisons, and judged writing quality | 106 of 187 | 48 of 187 |
| Sticking to the factsSummarising without inventing things, and factual answers | 116 of 153 | not measured |
Every shared result
| Benchmark | Metric | MiniMax M2.1 | Kimi K2 Thinking |
|---|---|---|---|
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 20.9 | 22.0 |
| Artificial Analysisreported | GPQA Diamond% of questions | 83.0% | 83.8% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 23.2% | 23.8% |
| Artificial Analysisreported | IFBench% of instructions | 69.9% | 68.1% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 67.7% | 72.0% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 28.8% | 31.1% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 85.4% | 93.0% |
| OpenHands Indexreported | Average% resolved | 41.2% | 41.0% |
| OpenHands Indexreported | Front end (SWE-Bench Multimodal)% resolved | 16.2% | 32.4% |
| OpenHands Indexreported | Greenfield (Commit0)% resolved | 18.8% | 12.5% |
| OpenHands Indexreported | Information gathering (GAIA)% resolved | 40.6% | 43.6% |
| OpenHands Indexreported | Issue resolution (SWE-Bench)% resolved | 68.8% | 69.2% |
| OpenHands Indexreported | Testing (SWT-Bench)% resolved | 61.4% | 47.3% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 31.0% | 20.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.41 | 0.38 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 41.6 | 59.3 |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.
Quick answers
Is MiniMax M2.1 better than Kimi K2 Thinking?
It depends on the task. Among models available today, on our benchmarks and licensed sources, MiniMax M2.1 ranks higher for agents and tool use; Kimi K2 Thinking ranks higher for reasoning and knowledge and writing.
Which is cheaper, MiniMax M2.1 or Kimi K2 Thinking?
MiniMax M2.1 costs $0.52 per million tokens (three input to one output), against $1.07 for Kimi K2 Thinking.
Run MiniMax M2.1 and Kimi K2 Thinking on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.