Side-by-side technical and economic comparison between Meta / Together's Llama 4 Maverick (400B) — Together AI and Mistral AI's Codestral (Current Generation).
• Llama 4 Maverick (400B) — Together AI: Classified as a Legacy Reference model (Deprecated on Together AI (Reference)).
| Metric | Llama 4 Maverick (400B) — Together AI | Codestral (Current Generation) | Advantage |
|---|---|---|---|
| Provider Organization | Meta / Together | Mistral AI | — |
| Standard Input / 1M | $0.27 | $0.30 | Llama 4 Maverick (400B) — Together AI (10% lower) |
| Cached Input / 1M | None | $0.03 | Codestral (Current Generation) lower |
| Output / 1M | $0.85 | $0.90 | Llama 4 Maverick (400B) — Together AI lower |
| Context Window | 1M | 256k | Llama 4 Maverick (400B) — Together AI (1M) |
| Max Generation Tokens | 16.4k | 8.2k | Llama 4 Maverick (400B) — Together AI |
| Tokenizer Family | Meta Llama 3/4 Tiktoken (128k vocabulary) | Mistral Tekken BPE (~131k vocabulary) | — |
Opt for Llama 4 Maverick (400B) — Together AI when your engineering requirements prioritize Meta / Together's ecosystem, specific tokenizer efficiencies (Calibrated Llama 3 Tokenizer (±3%)), or when your expected prompt-to-completion ratios favor its $0.27/M input rate.
Opt for Codestral (Current Generation) when looking for Mistral AI's tooling integration, specific context window depth (256k tokens), or when output generation volume favors its $0.9/M rate.