Compare / MiniMax M3 vs Inkling
MiniMax M3 vs Inkling
Which is cheaper, which is more capable, and which does better at each kind of work, from 23 results on 3 sources that measured both models.
MiniMax M3 or Inkling? The short answer
- MiniMax M3 is cheaper: $0.52 against $1.72 per million tokens, 3.3 times less.
- MiniMax M3 scores higher on the Artificial Analysis Intelligence Index: 29.2 against 25.0.
- Inkling writes faster: 166 tokens a second against 81.
- On our business benchmarks, MiniMax M3 ranks higher for professional work and reasoning and knowledge.
- On our business benchmarks, Inkling ranks higher for agents and tool use, writing and sticking to the facts.
At a glance
| MiniMax M3 | Inkling | |
|---|---|---|
| Price per million tokens | $0.30 in · $1.20 out | $0.95 in · $4.05 out |
| Blended (3 in : 1 out) | $0.52 | $1.72 |
| Cheapest host, blended | $0.41 (CoreWeave, FP4) | $1.72 (DeepInfra, FP8) |
| Intelligence Index | 29.2 | 25.0 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | 81 | 166 |
| Context window | 1,048,576 tokens | 524,288 tokens |
| Weights | Open | Open |
| Developer based in | China | the US |
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 M3 | Inkling |
|---|---|---|
| Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls | 71 of 161 | 56 of 161 |
| Professional workReal tasks from banking, consulting and law, business documents and freelance projects | 39 of 51 | 45 of 51 |
| Reasoning and knowledgeHard questions across science, maths and general knowledge | 36 of 195 | 64 of 195 |
| WritingWhat people prefer in blind comparisons, and judged writing quality | 61 of 187 | 59 of 187 |
| Sticking to the factsSummarising without inventing things, and factual answers | 94 of 153 | 75 of 153 |
Every shared result
| Benchmark | Metric | MiniMax M3 | Inkling |
|---|---|---|---|
| APEX-Agentsreported | Tasks passed% of tasks | 37.7% | 33.8% |
| Arena (formerly LMArena)reported | Confirmed task successIPS effect estimate | -0.14 | -0.22 |
| Arena (formerly LMArena)reported | Praise over complaintIPS effect estimate | -0.13 | -0.22 |
| Arena (formerly LMArena)reported | SteerabilityIPS effect estimate | -0.07 | -0.12 |
| Arena (formerly LMArena)reported | Tool groundingIPS effect estimate | -0.00 | -0.00 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,434 | 1,452 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,450 | 1,444 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,406 | 1,384 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,471 | 1,474 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,435 | 1,427 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,440 | 1,442 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,422 | 1,403 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 58.6 | 52.1 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 29.2 | 25.0 |
| Artificial Analysisreported | GPQA Diamond% of questions | 92.9% | 87.2% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 39.0% | 31.9% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 83.0% | 77.3% |
| Artificial Analysisreported | Output speedtokens per second | 80.6 | 166 |
| Artificial Analysisreported | SciCode% of problems | 47.1% | 47.0% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 65.2% | 55.1% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 2.0% | 1.0% |
| Artificial Analysisreported | Time to first answer tokenseconds | 25.9 | 14.3 |
| Artificial Analysisreported | Τ-bench banking% of tasks | 15.3% | 29.1% |
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 M3 better than Inkling?
It depends on the task. Among models available today, on our benchmarks and licensed sources, MiniMax M3 ranks higher for professional work and reasoning and knowledge; Inkling ranks higher for agents and tool use, writing and sticking to the facts.
Which is cheaper, MiniMax M3 or Inkling?
MiniMax M3 costs $0.52 per million tokens (three input to one output), against $1.72 for Inkling.
Which is faster, MiniMax M3 or Inkling?
Inkling writes about 166 tokens a second on its usual API, against 81 for MiniMax M3 (Artificial Analysis).
Run MiniMax M3 and Inkling on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.