Mistral AIMixtral 8x22B Instruct
Mixtral 8x22B Instruct is a Mistral AI open-weights model, released 17 Apr 2024. Its best configuration ranks #730 of 785 on the BenchLeader Index at 37.4 ±3.4, in the lower half. The ± is the point: 176 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (38) and lowest in coding (28). At $3.00 per million tokens blended it is pricier than most ranked models. Output speed of 104 tokens per second puts it faster than most, with a first token in 0.5 s. Last measured 22 Oct 2026.
- Blended price
- $3.00/M
- $2.00 in · $6.00 out
- Output speed
- 104 tok/s
- OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
- First answer
- 0.47 s
- Context
- 66k
- Full answer
- –
- Cost per run
- –
- Released
- 17 Apr 2024
- Overall index37
- Reasoning38
- Coding28
- Agents & tools37
- Maths35
- Knowledge29
- Instruction following34
- Human preference37
- Composite36
How its index has moved
9 Sept 2026 to 23 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Hard Prompts | 1244#294 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 33.2%#504 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 4.0%#551 | Artificial Analysis | |
| GPQA Diamond (HELM)not in index | 33.4%#57 | HELM Capabilities | |
| DTBenchnot in index | 55.1%#196 | Epoch AI Benchmarking Hub | |
| ForecastBenchnot in index | 58.8%#76 | ForecastBench |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| WeirdML | 3.2%#155 | WeirdML | |
| Arena Coding | 1278#290 | Arena (LMArena) |
Maths
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Omni-MATH (HELM) | 16.3%#59 | HELM Capabilities | |
| Arena Maths | 1228#279 | Arena (LMArena) |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| MedQA | 62.1%#82 | Vals AI | |
| MMLU-Pro (HELM)not in index | 46.0%#57 | HELM Capabilities | |
| MMLUnot in index | 77.8%#30 | Epoch AI Benchmarking Hub |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| IFEval (HELM) | 72.4%#63 | HELM Capabilities | |
| WildBench (HELM) | 71.1%#58 | HELM Capabilities | |
| Arena Instruction Followingnot in index | 1215#294 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1230#298 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1190#300 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1189#304 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1220#303 | Arena (LMArena) |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Epoch Capabilities Indexnot in index | 122.0#169 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 5.7#562 | Artificial Analysis | |
| HELM Capabilities meannot in index | 47.8%#58 | HELM Capabilities |
Safety & honesty
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| FORTRESSnot in index | 56.1%#8 | Scale AI SEAL |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Mistral | 104 tok/s | 0.44 s | $2.00 | $6.00 | 66k | – |
| Mistral | 97 tok/s | 0.53 s | $2.00 | $6.00 | 66k | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0026 | – | 3.4 s |
| Summarize a 30-page report | 12,000 / 600 | $0.028 | – | 6.3 s |
| Code edit | 6,000 / 1,500 | $0.021 | – | 15.0 s |
| Agentic coding session | 60,000 / 4,000 | $0.144 | – | 39.1 s |
| Structured extraction | 2,000 / 200 | $0.0052 | – | 2.4 s |
More from Mistral AI
See also
Questions people ask
- What are Mixtral 8x22B Instruct's benchmark scores?
- As of 25 Oct 2026, Mixtral 8x22B Instruct scores 37.4 ±3.4 on the BenchLeader Index, strongest in reasoning (38) and weakest in coding (28). That rests on 11 independent benchmark results, and every score on this page links to its publisher.
- How does Mixtral 8x22B Instruct rank?
- Mixtral 8x22B Instruct is #730 of 785 ranked configurations on 25 Oct 2026. 176 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Mixtral 8x22B Instruct cost per 1M tokens?
- As of 25 Oct 2026, Mixtral 8x22B Instruct costs $2.00 per 1M input tokens and $6.00 per 1M output tokens, $3.00 blended (three input tokens per output token).
- When was Mixtral 8x22B Instruct released?
- Mistral AI released Mixtral 8x22B Instruct on 17 Apr 2024.
Data as of 25 Oct 2026. Compare with another model.
Cite as: BenchLeader, “Mixtral 8x22B Instruct: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/mixtral-8x22b.