Model comparison
GPT-5 Nano vs Mixtral 8x22B
GPT-5 Nano is the stronger model overall, scoring 33.5 to 27.1 on the Noometry Index.
Last verified . 27 shared benchmarks.
Mistral AISummary
- They share 27 benchmarks with published results for both. GPT-5 Nano scores higher in 7 categories and Mixtral 8x22B in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-5 Nano leads 35.9 to 15.1.
- The biggest single-benchmark swing is MATH Level 5: 95.2% for GPT-5 Nano and 24.2% for Mixtral 8x22B.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $2 / $6 for Mixtral 8x22B.
- GPT-5 Nano accepts more context: 400K tokens versus 64K.
- Mixtral 8x22B has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Nano | Mixtral 8x22B | |
|---|---|---|
| Provider | OpenAI | Mistral AI |
| Noometry Index | 33.5 | 27.1 |
| Released | 2025-08-07 | 2024-04-17 |
| Weights | Proprietary | Open |
| Context window | 400K | 64K |
| Max output | 128K | 64K |
| Input $ / M tokens | $0.05 | $2 |
| Output $ / M tokens | $0.40 | $6 |
| Results tracked | 49 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 Nano leads
GPT-5 Nano: 33.6 (#254), Mixtral 8x22B: 24.2 (#329)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| WeirdML | 38.1% | 3.2% |
| LMArena Coding | 1351 | 1166 |
| SWE-bench Verified (bash only) | 34.8% | — |
| BigCodeBench Instruct | — | 40.6% |
| BigCodeBench Complete | — | 50.2% |
| ALE-Bench | 718.67 | — |
| HumanEval+ | — | 72% |
| MBPP+ | — | 64.3% |
Agentic & Tool Use GPT-5 Nano leads
GPT-5 Nano: 25.8 (#106), Mixtral 8x22B: 23.1 (#127)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| Terminal-Bench | 21.8% | — |
| Berkeley Function Calling Leaderboard | 51.5% | — |
| Cybench | — | 7.5% |
Reasoning Mixtral 8x22B leads
GPT-5 Nano: 16.3 (#306), Mixtral 8x22B: 19.9 (#248)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| LMArena Hard Prompts | 1328 | 1150 |
| DTBench | 62.7% | 55.1% |
| Epoch Capabilities Index | 139.38 | 122.03 |
| ForecastBench | 59.1 | 56.3 |
| ARC-AGI-2 | 2.6% | — |
| Kagi LLM Benchmark | 62.2% | — |
| ARC-AGI-1 | 20.7% | — |
| Chess Puzzles | 27% | — |
| Mystery Game Puzzles | 9% | — |
| LMCA | 7.9% | — |
Math GPT-5 Nano leads
GPT-5 Nano: 29.4 (#241), Mixtral 8x22B: 22.9 (#275)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| Omni-MATH | 54.6% | 16.3% |
| LMArena Math | 1317 | 1184 |
| MATH Level 5 | 95.2% | 24.2% |
| FrontierMath (Tiers 1-3) | 20% | — |
| FrontierMath Tier 4 | 2.4% | — |
| OTIS Mock AIME 2024-2025 | 81.1% | — |
| ProofBench | 12% | — |
| FrontierMath (Feb 2025 set) | 8.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-5 Nano leads
GPT-5 Nano: 35.9 (#178), Mixtral 8x22B: 15.1 (#293)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| GPQA Diamond | 69.4% | 34.1% |
| MMLU-Pro | 77.8% | 46% |
| GPQA (HELM) | 67.9% | 33.4% |
| LMArena Expert | 1321 | 1113 |
| SimpleQA Verified | 11.7% | — |
| Vectara Hallucination Rate | 10.5% | — |
| MMLU | — | 77.8% |
Multimodal Not comparable
GPT-5 Nano: 31.3 (#108), Mixtral 8x22B: —
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| LMArena Vision | 1159 | — |
| VPCT | 37.2% | — |
Multilingual GPT-5 Nano leads
GPT-5 Nano: 45.3 (#172), Mixtral 8x22B: 32.8 (#255)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| LMArena Non-English | 1313 | 1128 |
| LMArena Chinese | 1356 | 1116 |
| LMArena German | 1327 | 1141 |
| LMArena Japanese | 1226 | 1037 |
| LMArena Korean | 1269 | 1057 |
| LMArena Russian | 1296 | 1158 |
| LMArena Spanish | 1360 | 1151 |
| LMArena French | — | 1166 |
Instruction Following GPT-5 Nano leads
GPT-5 Nano: 75.0 (#79), Mixtral 8x22B: 57.7 (#266)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| IFEval | 93.2% | 72.4% |
| LMArena Instruction Following | 1306 | 1147 |
Long Context Mixtral 8x22B leads
GPT-5 Nano: 31.3 (#281), Mixtral 8x22B: 34.7 (#247)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| LMArena Longer Query | 1312 | 1144 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference GPT-5 Nano leads
GPT-5 Nano: 39.1 (#249), Mixtral 8x22B: 36.9 (#262)
| Benchmark | GPT-5 Nano | Mixtral 8x22B |
|---|---|---|
| LMArena Text | 1320 | 1162 |
| LMArena Creative Writing | 1249 | 1141 |
| WildBench | 80.6% | 71.1% |
| LMArena Multi-Turn | 1311 | 1130 |
| EQ-Bench Creative Writing | 705 | — |
Frequently asked questions
Is GPT-5 Nano better than Mixtral 8x22B?
GPT-5 Nano is the stronger model overall, scoring 33.5 to 27.1 on the Noometry Index.
Which is cheaper, GPT-5 Nano or Mixtral 8x22B?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Mixtral 8x22B lists at $2 and $6.
Is GPT-5 Nano or Mixtral 8x22B better for coding?
GPT-5 Nano scores higher on coding benchmarks: 33.6 versus 24.2 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 64K.
How many benchmarks do GPT-5 Nano and Mixtral 8x22B share?
27 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and Mixtral 8x22B has 34.