Mistral AILanguage model
Pixtral Large (latest)
Pixtral Large (latest) has no overall index value yet because there are too few independent benchmark results.
PopularityRank 689 of 785
22out of 100-1 vs. last month
based on search, usage and attention
Summary
Pixtral Large (latest) is the large multimodal Pixtral checkpoint with a 123B decoder and a vision encoder of about 1B parameters. It accepts text and images with 128,000 tokens of context and, according to Mistral, can handle at least 30 high-resolution images in one context. Documents, diagrams, and complex image questions are its focus.
Mistral introduced Pixtral Large as a multimodal extension of Mistral Large 2.1. The Mistral Research License covers research, and the current overview lists the rolling alias as deprecated.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 20 published benchmark results
Page 1 of 2
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
TurtleAI DSReal (Show measurement, test conditions, and source)
- Source value
- 10.78
- Score
- 10.78
- Metric
- Symbolic success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 102
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DSReal (Show measurement, test conditions, and source)
- Source value
- 11.76
- Score
- 11.76
- Metric
- Embedding success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 102
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DSCraft (Show measurement, test conditions, and source)
- Source value
- 13.73
- Score
- 13.73
- Metric
- Symbolic success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 102
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DSCraft (Show measurement, test conditions, and source)
- Source value
- 15.69
- Score
- 15.69
- Metric
- Embedding success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 102
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DSSyn (Show measurement, test conditions, and source)
- Source value
- 4.68
- Score
- 4.68
- Metric
- Symbolic success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 619
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DSSyn (Show measurement, test conditions, and source)
- Source value
- 3.88
- Score
- 3.88
- Metric
- Embedding success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 619
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DS combined (Show measurement, test conditions, and source)
- Source value
- 6.56
- Score
- 6.56
- Metric
- Symbolic success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 823
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DS combined (Show measurement, test conditions, and source)
- Source value
- 6.32
- Score
- 6.32
- Metric
- Embedding success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 1; greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 823
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DSReal, chain-of-thought prompt (Show measurement, test conditions, and source)
- Source value
- 11.76
- Score
- 11.76
- Metric
- symbolic success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 5; CoT prompt, greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 102
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
TurtleAI DS combined, chain-of-thought prompt (Show measurement, test conditions, and source)
- Source value
- 4.74
- Score
- 4.74
- Metric
- symbolic success
- Unit
- %
- Benchmark version
- TurtleAI, Findings of ACL 2026 Table 5; CoT prompt, greedy decoding
- Category
- visual-programming
- Direction
- Higher is better
- Sample
- 823
- Source type
- independent-evaluator
- Evaluator
- TurtleAI paper, Findings of ACL 2026
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Exact Hugging Face checkpoint mistralai/Pixtral-Large-Instruct-2411, cross-checked against the official API alias pixtral-large-2411. TurtleAI evaluates visual programming: provide task image, generate Python turtle code, then compare generated output with reference. Main results use greedy decoding; Table 5 CoT row is a separate prompt condition. Dataset N: DSReal 102, DSCraft 102, DSSyn 619, combined DS 823.
- Context
- No confidence intervals are published. Table 1 main-result rows that duplicate the no-CoT Table 5 values are kept only once; the separate Table 5 CoT rows remain below.
HumanEval-V visual coding, pass@1 (Show measurement, test conditions, and source)
- Source value
- 11.1
- Score
- 11.1
- Metric
- pass@1
- Unit
- %
- Benchmark version
- HumanEval-V, 2024-11-20 official result
- Category
- vision-coding
- Direction
- Higher is better
- Sample
- 108
- Source type
- official-leaderboard
- Evaluator
- HumanEval-V official benchmark repository, README at commit 05d4b520fe41807e2d1b10355c2be77feb6eb08c
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact checkpoint: Pixtral-Large-Instruct-2411. The authors' 2024-11-20 announcement reports pass@1 11.1. The repository's general protocol uses greedy decoding for pass@1. The result is a historical checkpoint measurement, not a statement about the current pixtral-large-latest API target.
- Context
- The dataset contains 108 tasks. The repository gives neither a confidence interval nor model-specific hardware, seed, or decoding details beyond the general pass@1 rule. The result cannot be transferred to a current API alias.
HumanEval-V visual coding, pass@10 (Show measurement, test conditions, and source)
- Source value
- 26.9
- Score
- 26.9
- Metric
- pass@10
- Unit
- %
- Benchmark version
- HumanEval-V, 2024-11-20 official result
- Category
- vision-coding
- Direction
- Higher is better
- Sample
- 108
- Source type
- official-leaderboard
- Evaluator
- HumanEval-V official benchmark repository, README at commit 05d4b520fe41807e2d1b10355c2be77feb6eb08c
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Official benchmark leaderboard. Exact checkpoint: Pixtral-Large-Instruct-2411. The authors' 2024-11-20 announcement reports pass@10 26.9. The repository's general protocol samples 20 generations at temperature 0.8 and top-p 0.95 for pass@10. The result is a historical checkpoint measurement, not a statement about the current pixtral-large-latest API target.
- Context
- The dataset contains 108 tasks. The 20 generations per task are not 20 independent test tasks. The repository gives no confidence interval or model-specific hardware or seed details. The result cannot be transferred to a current API alias.
Page 1 of 2
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
The 100 most popular models of this category, 0 to 100, higher is more popular
20 of 101 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 7, 2026
Pixtral Large (latest)MiniMax M2.5 22Grok 4.3 22InternLM2-Chat-20B 22Command A Reasoning 21
Method changes in the timeline, a jump at these points is no market event:
- 1November 1, 2026: Value only with search or usage data, 30-day attention
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Similar models
More models of the same kind as Pixtral Large (latest), chosen by model line, developer, size, price and release date.
- Mistral Large 2.1Mistral AI · same developer
- Mistral Large 2407Mistral AI · same developer
- Mixtral 8x22BMistral AI · same developer
- Devstral 2Mistral AI · same developer
- Mistral Small 4Mistral AI · same developer
- Mistral Medium 3.5Mistral AI · same developer
- Magistral SmallMistral AI · same developer
- Devstral SmallMistral AI · same developer
All specifications
Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
Frequently asked questions about Pixtral Large (latest)
What is the knowledge cutoff of Pixtral Large (latest)?
November 2024, according to Mistral AI. That is the month of release.
Stated by Mistral AI
NovKnowledge cutoffNovember 2024ReleasedNov 1, 2024
How large is the context window of Pixtral Large (latest)?
128,000 tokens. That is how much text Pixtral Large (latest) reads at once, about 190 A4 pages.
Page count as a rule of thumb from Artificial Analysis
GLM-5.1200K
DeepSeek-V3 0324164K
GLM-4.5131K
GPT OSS 120B131K
Pixtral Large (latest)Rank 98 of 103128K
DeepSeek-V3.2128K
DeepSeek-R1128K
Can I run Pixtral Large (latest) locally?
Yes, the weights are open. With 124B parameters it needs a lot of memory, more a server than a laptop.
Stated by Mistral AI
- Open weights
- yes
- License
- Mistral Research License
- Parameters
- 124B
- active per token
- n/a
Sources and data date
Show all 15 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
No active result from the reviewed external evaluators can currently be attributed to this model. Other source evidence remains available below.
Without vendor claims and historical or disputed results; those stay labelled in the single values.
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved October 27, 2026. |
| Output | Documented modalities. Retrieved October 27, 2026. |
| Research status | Research date October 15, 2026. 6 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Unresolved | current-api-alias Note The release proves the historical alias-to-2411 checkpoint mapping; after the publisher deprecated the model, no current alias target is claimed. |
| Not published | benchmark-configuration Note The model card gives benchmark names and metric labels but omits run dates, prompts, sample counts and decoding settings. |
| Not published | confidence-intervals Original wording TurtleAI result tables do not publish model-level confidence intervals. |
| Additional source | Hugging Face (retrieved October 15, 2026) · Pixtral Large model card · Editorial description reviewed October 2, 2026 |
| Additional source | TurtleAI paper, Findings of ACL 2026 (retrieved October 17, 2026) |
| Additional source | HumanEval-V official benchmark repository, README at commit 05d4b520fe41807e2d1b10355c2be77feb6eb08c (retrieved October 17, 2026) |
| Additional source | BenchLeader, Pixtral Large model page and Index methodology (retrieved October 19, 2026) |
| Additional source | Model metadata (retrieved October 15, 2026) |
| Additional source | Model metadata (retrieved October 17, 2026) |
| Additional source | Pixtral Large announcement · Editorial description reviewed October 2, 2026 |
| Additional source | Mistral model overview · Editorial description reviewed October 2, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Ollama: cumulative downloads for the popularity index |
| Additional source | Replicate: cumulative runs for the popularity index |
| Additional source | Google Trending Now: limited feed selection for Germany and the US |
| Additional source | Creative Commons Attribution 4.0 |