Compare / Mistral Small 3.2 24B vs GPT-5.6 Terra
Mistral Small 3.2 24B vs GPT-5.6 Terra
Which is cheaper, which is more capable, and which does better at each kind of work, from 21 results on 3 sources that measured both models.
Mistral Small 3.2 24B or GPT-5.6 Terra? The short answer
- Mistral Small 3.2 24B is cheaper: $0.13 against $4.50 per million tokens, 34 times less.
- GPT-5.6 Terra scores higher on the Artificial Analysis Intelligence Index: 42.1 against 8.2.
- On our business benchmarks, GPT-5.6 Terra ranks higher for agents and tool use, reasoning and knowledge and writing.
At a glance
| Mistral Small 3.2 24B | GPT-5.6 Terra | |
|---|---|---|
| Price per million tokens | $0.094 in · $0.25 out | $2.00 in · $12.00 out |
| Blended (3 in : 1 out) | $0.13 | $4.50 |
| Cheapest host, blended | $0.11 (DeepInfra, FP8) | $2.25 (OpenAI) |
| Intelligence Index | 8.2 | 42.1 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | – | 104 |
| Context window | 256,000 tokens | 1,050,000 tokens |
| Weights | Open | Closed |
| Developer based in | France | 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 | Mistral Small 3.2 24B | GPT-5.6 Terra |
|---|---|---|
| Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls | 118 of 161 | 16 of 161 |
| Professional workReal tasks from banking, consulting and law, business documents and freelance projects | not measured | 10 of 51 |
| Reasoning and knowledgeHard questions across science, maths and general knowledge | 163 of 195 | 20 of 195 |
| WritingWhat people prefer in blind comparisons, and judged writing quality | 130 of 187 | 39 of 187 |
| Sticking to the factsSummarising without inventing things, and factual answers | not measured | 49 of 153 |
Every shared result
| Benchmark | Metric | Mistral Small 3.2 24B | GPT-5.6 Terra |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,363 | 1,466 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,323 | 1,428 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,335 | 1,504 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,338 | 1,462 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,357 | 1,465 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,329 | 1,445 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 12.5 | 76.7 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 8.20 | 42.1 |
| Artificial Analysisreported | GPQA Diamond% of questions | 50.5% | 92.5% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 4.3% | 42.9% |
| Artificial Analysisreported | IFBench% of instructions | 33.5% | 71.2% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 20.3% | 83.0% |
| Artificial Analysisreported | SciCode% of problems | 28.6% | 55.0% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 5.6% | 88.0% |
| Artificial Analysisreported | Terminal-Bench 4.0% of tasks | 0.0% | 35.4% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 6.8% | 62.9% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 6.2% | 40.2% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 29.5% | 86.3% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 17.0% | 22.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.35 | 0.39 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 36.3 | 60.4 |
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 Mistral Small 3.2 24B better than GPT-5.6 Terra?
It depends on the task. Among models available today, on our benchmarks and licensed sources, GPT-5.6 Terra ranks higher for agents and tool use, reasoning and knowledge and writing.
Which is cheaper, Mistral Small 3.2 24B or GPT-5.6 Terra?
Mistral Small 3.2 24B costs $0.13 per million tokens (three input to one output), against $4.50 for GPT-5.6 Terra.
Run Mistral Small 3.2 24B and GPT-5.6 Terra on your own prompt
Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.