Model comparison
Mistral Small vs Sonar
Sonar is the stronger model overall, scoring 38.5 to 33.4 on the Noometry Index. Mistral Small costs 3.8× less per token, which makes it the better buy when Sonar's lead doesn't matter for your workload.
Last verified . 7 shared benchmarks.
Mistral AI
PerplexitySummary
- They share 7 benchmarks with published results for both. Mistral Small scores higher in 0 categories and Sonar in 5 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where Sonar leads 33.7 to 16.4.
- The biggest single-benchmark swing is LiveBench Data Analysis: 53.7% for Mistral Small and 37.9% for Sonar.
- Mistral Small is cheaper at $0.15 / $0.60 per million input/output tokens, against $1 / $1 for Sonar.
- Mistral Small accepts more context: 262K tokens versus 128K.
- Mistral Small has downloadable open weights; the other is API-only.
Side by side
| Mistral Small | Sonar | |
|---|---|---|
| Provider | Mistral AI | Perplexity |
| Noometry Index | 33.4 | 38.5 |
| Released | 2024-02-26 | 2024-01-01 |
| Weights | Open | Proprietary |
| Context window | 262K | 128K |
| Max output | 256K | 4K |
| Input $ / M tokens | $0.15 | $1 |
| Output $ / M tokens | $0.60 | $1 |
| Results tracked | 39 | 7 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Sonar leads
Mistral Small: 34.0 (#247), Sonar: 35.7 (#221)
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LiveBench Coding | 36.2% | 35.1% |
| SciCode | 26.5% | — |
| BigCodeBench Instruct | 36.1% | — |
| LMArena Coding | 1362 | — |
| BigCodeBench Complete | 46.6% | — |
| ALE-Bench | 497.62 | — |
Agentic & Tool Use Not comparable
Mistral Small: 28.1 (#93), Sonar: —
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| Berkeley Function Calling Leaderboard | 37.1% | — |
Reasoning Sonar leads
Mistral Small: 19.8 (#250), Sonar: 21.1 (#227)
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LiveBench Reasoning | 44.8% | 46.3% |
| LiveBench Data Analysis | 53.7% | 37.9% |
| LiveBench | 44% | 46.9% |
| Kagi LLM Benchmark | 37.8% | — |
| CritPt | 0% | — |
| LMArena Hard Prompts | 1335 | — |
| DTBench | 70.9% | — |
| LMCA | 20.6% | — |
Math Sonar leads
Mistral Small: 16.4 (#293), Sonar: 33.7 (#200)
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LiveBench Math | 39.9% | 41.6% |
| OTIS Mock AIME 2024-2025 | 5.8% | — |
| LMArena Math | 1341 | — |
| MATH Level 5 | 46.8% | — |
Knowledge Not comparable
Mistral Small: 31.0 (#222), Sonar: —
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| GPQA Diamond | 47.5% | — |
| Vectara Hallucination Rate | 5.1% | — |
| LMArena Expert | 1291 | — |
| MMLU | 68.7% | — |
Multimodal Not comparable
Mistral Small: 33.5 (#96), Sonar: —
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LMArena Vision | 1142 | — |
Multilingual Not comparable
Mistral Small: 45.5 (#169), Sonar: —
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LMArena Non-English | 1315 | — |
| LMArena Chinese | 1340 | — |
| LMArena French | 1337 | — |
| LMArena German | 1340 | — |
| LMArena Japanese | 1275 | — |
| LMArena Korean | 1259 | — |
| LMArena Russian | 1324 | — |
| LMArena Spanish | 1346 | — |
Instruction Following Sonar leads
Mistral Small: 66.4 (#209), Sonar: 71.4 (#150)
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LiveBench Instruction Following | 63.7% | 76.2% |
| LMArena Instruction Following | 1310 | — |
Long Context Not comparable
Mistral Small: 40.4 (#156), Sonar: —
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LMArena Longer Query | 1327 | — |
Writing & Preference Too close to call
Mistral Small: 52.5 (#171), Sonar: 52.6 (#167)
| Benchmark | Mistral Small | Sonar |
|---|---|---|
| LiveBench Language | 30.5% | 44.1% |
| LMArena Text | 1338 | — |
| LMArena Creative Writing | 1305 | — |
| LMArena Multi-Turn | 1344 | — |
Frequently asked questions
Is Mistral Small better than Sonar?
Sonar is the stronger model overall, scoring 38.5 to 33.4 on the Noometry Index. Mistral Small costs 3.8× less per token, which makes it the better buy when Sonar's lead doesn't matter for your workload.
Which is cheaper, Mistral Small or Sonar?
Mistral Small is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Sonar lists at $1 and $1.
Is Mistral Small or Sonar better for coding?
Sonar scores higher on coding benchmarks: 35.7 versus 34.0 in the Noometry coding category.
Which has the bigger context window?
Mistral Small does, with 262K tokens against 128K.
How many benchmarks do Mistral Small and Sonar share?
7 benchmarks have published results for both models. Mistral Small has 39 scored results on Noometry and Sonar has 7.