Mixtral 8x22B Instruct
Mistral AIReleased Apr 17, 2024Open weights · Apache 2.0mixtral-8x22b
- Intelligence (Epoch)
- 122 · 105th of 111
- Reasoning
- no
- Context
- 66K tokens
- Max output
- 52k tokens
- Input
- text
- API features
- tools, structured outputs, JSON mode, seed
- Mistral AI's API
- yes, as open-mixtral-8x22b
Read Oct 23, 06:00 UTC from the hosts' published offers and OpenRouter, spec sheet from Mistral AI
Intelligence for the price
81 models score higher for the same price or less; it ranks 82 of the 88 models at its price or below.
Epoch Capabilities Index against the lowest input + output price per M, log scale · prices Oct 23. Score: Epoch AI, CC BY 4.0, 122 with an interval of 115 to 124.
Against its rivals
The best of its price range and the top score
- Mixtral 8x22B Instructthis model$8.00122
- Claude Opus 5.5top score$24.00169
- GPT-6.1 Solsame price range$12.00168
- Claude Sonnet 5.5same price range$12.00166
Same level for 211× less: Mistral Nemo at Io Net, $0.038 per M input + output · their Epoch intervals overlap (119 against 122).
Bars: Epoch Capabilities Index (Epoch AI, CC BY 4.0), drawn from 100, below the lowest score in the catalogue · price: lowest input + output per M.
Run it yourself
open weights, Apache 2.0 · fits 4 of the 6 workstations in our guideSelf-host guide ›- Parameters
- 141B
- 39B active per token (MoE)
- Memory needed
- 85 GB
- in INT4 · 169 GB in INT8
- Smallest cloud setup
- 2× H100
- 160 GB · $4.30/h, about $3,139 a month
- Worth it above
- 392M tokens
- a month, against the cheapest API at $8.00 per M
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 85 GB | 2× H100 · $4.30/h | smallest footprint, minor quality loss |
| INT8 (Q8) | 169 GB | 4× H100 · $8.60/h | near-lossless |
| FP16 (full) | 338 GB | 8× H100 · $17.20/h | reference quality |
At home: in INT4 it fits a Mac Studio M5 Ultra (96 GB unified), about $5,499 once, or $0.2092/h over 3 years, electricity aside; it pays for itself after about 687M tokens against the cheapest API.
Memory ≈ parameters × bytes per parameter × 1.2 · cloud price: on-demand rate, running 24/7 · builds: GGUF, AWQ, MLX, GPTQ, Ollama · break-even calculator
| Provider | Input | Output | Cache ↑ | tok/s | First token | Uptime 24h | Checked |
|---|---|---|---|---|---|---|---|
| Mistral AI64K ctx · 52K out | $2.00 | $6.00 | $0.20 | 196 | 0.43 s | 100% | Oct 23 |
1 provider · prices per million tokens, read Oct 23, 06:00 UTC · speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · uptime: OpenRouter, last 24 h · under each host: weights precision, context and max output as served, the maker's value in brackets when the host's differs
| Provider | Weights | Context | Max output | Tools | Structured outputs | Reasoning |
|---|---|---|---|---|---|---|
| Mistral AI | not checked yet | 64K | 52K | yes | yes | no |
Weights, context and max output: read at each host's own catalogue or docs where its terms allow, otherwise as the host declares to OpenRouter; hover a cell for its source and date · tools, structured outputs, reasoning: as declared to OpenRouter · not verified by us
Mixtral 8x22B Instruct spec sheet, field by field, with sources
Mixtral 8x22B Instruct spec sheetmaker-declared
What Mistral AI declares about the model itself — read from its own weights repository and documentation, not from any host. What each endpoint actually serves is in the pricing above: the two can differ, and both are true.
- Context window
- 65,536 tokens
- Parameters
- 140,630,071,296≈ 140.6B
- Active experts
- 2 of 8 active per token
- Weight precision
- bfloat16
- License
- Apache 2.0 — open weights, permissivecurated · read 2026-09-07Scope: the licence on the published weights, and nothing else. It is not the maker's acceptable-use policy, which is a separate document, and it is not the contract you sign with whichever host you call — hosts set their own terms, which this index does not read.
- Knowledge cutoff
- not published by the makermaker's docs · read 2026-09-07
- Modalities
- text
- Max output (official)
- to verify
- Release date
- April 17, 2024maker's docs · read 2026-09-07
- Maker lifecycle status
- Retired — the maker withdrew it on March 30, 2025 · superseded by Mistral Small 4maker's docs · read 2026-09-07
Also declared in the weights
- Architecture class
- MixtralForCausalLM
- Layers
- 56
- Vocabulary
- 32,768 tokens
- Weights first published
- April 16, 2024
- Weights last updated
- July 24, 2025
- Languages declared
- English · Spanish · Italian · German · French
mistralai/Mixtral-8x22B-Instruct-v0.1 · config.json @ cc88a6c · read 2026-10-08 · 8 of 10 fields published by the maker
| Date | Provider | Input | Output | Change |
|---|---|---|---|---|
| Sep 5 | Mistral AI | $2.20 → $2.00 | $6.60 → $6.00 | −9% |
Every published change of input or output price, newest first · change on input + output · full series: pricing-history.csv.
| Provider | Chatbot 1,000 in / 500 out | RAG / long context 8,000 in / 500 out | Batch summarize 4,000 in / 1,000 out |
|---|---|---|---|
| Mistral AI | $5.00 | $19.00 | $14.00 |
What each host bills for 1,000 requests of a representative job
Your own workload
Other Mistral AI models · all its models, hosts and history ›
- Mistral Nemo6 providers · from $0.038
- Mistral Small 3.2 24B4 providers · from $0.28
- Mistral Small 42 providers · from $0.75
- Codestral 25081 provider · from $1.20
- Devstral 2 25121 provider · from $2.40
- Ministral 3 14B 25121 provider · from $0.40