Llama 4 Maverick
- Context
- 1.0M
- Released
- 1y ago
- Parameters
- —
- Input / 1M tokens
- $0.19
- Output / 1M tokens
- $0.65
- Cached input / 1M tokens
- $0.05
Sources:OpenRouter
Benchmarks
DeepSWE is the primary coding benchmark where an exact model evaluation exists. Other sourced evaluations appear below; missing scores are not zero.
Not evaluated in DeepSWE 1.1 as of 2026-09-22. Results are only shown for exact model versions in the official dataset. Official leaderboard.
No benchmark results have been matched to this model yet.
Specifications
- Developer
- Meta
- Model family
- Llama
- Type
- Language model
- Predecessor
- —
- Successor
- —
- Released
- April 5, 2025
- Model ID
meta-llama/llama-4-maverick- Openness
- Open weights
- License
- —
- Parameters
- —
- Input modalities
- text, image
- Output modalities
- text
- Modality
- text+image->text
- Tokenizer
- Llama4
- Max output tokens
- 16K
- Knowledge cutoff
- 2024-08-31
- Hugging Face
- meta-llama/Llama-4-Maverick-17B-128E-Instruct
- Source freshness
- Endpoints synced Oct 15, 2026, 12:25 PM UTC
Infrastructure
Provider availability
| Provider | Endpoint | Context | Input /M tokens | Output /M tokens | Cached input /M tokens | Cache write /M tokens | Uptime 1d | Data handling |
|---|---|---|---|---|---|---|---|---|
| digitalocean | 128K | $0.19 | $0.65 | $0.05 | — | 98.89% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| novita/fp8 | 1.0M | $0.27 | $0.85 | — | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| google-vertex/us-east5 | 524K | $0.35 | $1.15 | — | — | — | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| parasail/fp8 | 524K | $0.35 | $1.00 | $0.17 | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 |
Provider data:OpenRouter
Deployment graph
Can I run this?
Vertical Tabs only shows compatibility when it has structured evidence tying this exact model to hardware, runtime, quantization, memory use, and observed performance.