Home › Frontier and closed › mistralai
Mistral Large 4 mistralai
The Hugging Face repo mistralai/Mistral-Large-4.0-1T05-A52B exists but is private: weights are expected, not yet public. Observed 2026-10-24 21:30 UTC.
Release timeline
Stages observed by the nightly radar: announced → API only → weights public → served by third parties. Each entry carries the source it was seen at.
- 2026-10-06 · API only source seeded from the earliest dated observation when first seen · observed 2026-10-07 20:20 UTC
- next · Next expected: weights on Hugging Face
Description
Mistral Large 4 is a frontier multimodal (text and image input) model from Mistral AI built for reasoning, coding, and agentic workloads. It offers a 1M-token context window with up...
via OpenRouter listing · observed 2026-10-24 21:26 UTC. The listing text is the provider's own description, not an assessment.
First look measured
A fixed probe set run against a live API endpoint: the same questions for every model, answers graded by code, nothing judged by a model. Latencies are single-request wall clock from one client location, not a benchmark.
| Probe | Result | Latency | Reply (truncated) |
|---|---|---|---|
| Self-description | info | 5.3 s · thinks | I am Mistral Large 4, created by Mistral AI. I don't know my exact training data cutoff date. |
| Tool call | pass | 2.0 s · thinks | {"city": "Lisbon", "unit": "c"} |
| Strict JSON | pass | 4.7 s · thinks | {"name": "Foo-9B-Instruct", "license": "Apache-2.0", "parameters_b": 9.2, "context_tokens": 131072, "languages": ["English", "French", "German", "Japanese"]} |
| Code (hidden tests) | pass | 10.6 s · thinks | def is_balanced(s: str) -> bool: stack = [] pairs = {')': '(', ']': '[', '}': '{'} for c in s: if c in '([{': stack.append(c) elif c in ')]}': if not stack or stack.pop() != pairs[c]: return False return not stack |
| Multi-step arithmetic | pass | 20.3 s · thinks | 16 |
| Needle recall at 128k tokens | pass | 14.7 s · thinks | amber-falcon-7731 |
Probe set firstlook.v1 · observed 2026-10-07 18:13 UTC · endpoint · provider docs · cost of this run $0.10
Where to run 2 providers · 4 endpoints
Hosted endpoints observed for Mistral Large 4, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $1.36/M.
| Provider | Model ref | Context | Quant | $/M in | $/M out | tok/s | TTFT | Tools | Status | Observed |
|---|---|---|---|---|---|---|---|---|---|---|
| Mistral openrouter | mistralai/mistral-large-4-0 | 1M | unknown | $1.36 | $4.18 | — | — | yes | live | 2026-10-24 |
| Mistral openrouter | mistralai/mistral-large-4-0 | 1M | unknown | $1.36 | $4.18 | — | — | yes | live | 2026-10-24 |
| Mistral openrouter | mistralai/mistral-large-4-0 | 512K | unknown | $1.50 | $4.60 | — | — | yes | live | 2026-10-24 |
| Mistral (first-party) mistral-api | mistral-large-4-0 | 1M | — | — | — | — | — | — | live | 2026-10-24 |
No third-party providers observed beyond the lab's own endpoints. Cache-read price $0.14/M (source, observed 2026-10-24).
Facts
| Context length | 1,048,576 tokens (1M) · OpenRouter listing |
|---|---|
| Modalities | text + image → text |
| Max completion tokens | 262,144 |
| Lab API ids | Mistral: mistral-large-4-0 |
| Supported parameters | frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, reasoning_effort, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_logprobs, top_p |
| Openness | Weights pending · HF repo exists but is private |
| Stage | API only |
| Hugging Face repo | mistralai/Mistral-Large-4.0-1T05-A52B (private) |
| Confidence | high |
| Pricing (USD per M tokens) |
in $1.36 · out $4.18 · cache read $0.14 OpenRouter list price (cheapest live endpoint) · source · observed 2026-10-24 21:26 UTC |
Open alternatives 5
Nearest open-weights models by modality, context and price on this watch list. Not a quality ranking.
| Model | Org | Params | Context | $/M in | License | Why it is here |
|---|---|---|---|---|---|---|
| Kimi-K3 | moonshotai | 2.78T | 1M | $0.23 | other | similar context (1.05M), cheaper input price ($0.2261/M vs $1.36/M), vision input |
| MiniMax-M3 | MiniMaxAI | 427B | 1M | $0.22 | other | similar context (1.05M), cheaper input price ($0.224/M vs $1.36/M), vision input |
| Qwen3.5-4B | Qwen | 4.66B | 256K | — | apache-2.0 | vision input |
| DeepSeek-V4.1-Flash-1017 | deepseek-ai | 763B | 1M | $0.20 | mit | similar context (1.05M), cheaper input price ($0.2/M vs $1.36/M), vision input |
| Kimi-K2.6 | moonshotai | 1.03T | 256K | $0.45 | other | cheaper input price ($0.44935/M vs $1.36/M), vision input |
Get alerts
Watch the release radar
- /feeds/upcoming.xml(RSS 2.0) · upcoming and just-released models, one item per entry
- /feeds/upcoming.json(JSON Feed 1.1) · same, JSON Feed 1.1
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.
Sources
- openrouter mistralai/mistral-large-4-0 · observed 2026-10-24 21:26 UTC
- mistral mistral-large-4-0 · observed 2026-10-24 21:26 UTC
- hf mistralai/Mistral-Large-4.0-1T05-A52B · observed 2026-10-24 21:30 UTC
Radar run: 2026-10-25 06:51 UTC · Last verified 2026-10-25 06:51 UTC