Codestral vs DeepSeek-Coder-V2 236B
Benchmarks, API Token Pricing, Context Window & Speed Comparison
Related topics (12)
Codestral di Mistral AI e DeepSeek-Coder-V2 236B di DeepSeek sono due modelli di IA di riferimento. Codestral offre una finestra di contesto di 256K token a $0.20/$0.60 per milione di token, mentre DeepSeek-Coder-V2 236B mette a disposizione 128K token di contesto a $0.14/$0.28 per milione di token. Di seguito analizziamo benchmark, latenza e costi.
Cambia e Confronta Qualsiasi Modello
Select any AI models from our verified benchmark database to generate live side-by-side matrices.
Codestral
Specialista Codice Open-WeightsCodestral di Mistral AI è un modello di IA progettato per lo sviluppo software e la generazione di codice con pesi aperti auto-ospitabili, con finestra di contesto da 256K, velocità di 110 token/s, indice di intelligenza di 40/100 e tariffa
$0.20 / 1M
$0.60 / 1M
256K tokens
28%
DeepSeek-Coder-V2 236B
Modello Base a Pesi ApertiDeepSeek-Coder-V2 236B di DeepSeek è un modello di IA progettato per il self-hosting su infrastruttura privata, il fine-tuning e la piena sovranità dei dati, con finestra di contesto da 128K, velocità di 74 token/s, indice di intelligenza d
$0.14 / 1M
$0.28 / 1M
128K tokens
42.1%
Differenze Chiave a Colpo d’Occhio
- DeepSeek-Coder-V2 236B is 1.9x cheaper for API usage (input + output tokens combined).
- Codestral has a larger context window (256K vs 128K tokens).
- Codestral is faster at 110 tok/s vs 74 tok/s.
- DeepSeek-Coder-V2 236B leads on SWE-bench coding (42.1% vs 28.0%).
| Feature | Codestral | DeepSeek-Coder-V2 236B |
|---|---|---|
Intelligence Index Composite reasoning & knowledge score (0-100) | 40 | 42 |
Provider | Mistral AI | DeepSeek |
Input Price Cost per million input tokens | $0.20 / 1M | $0.14 / 1M |
Output Price Cost per million output tokens | $0.60 / 1M | $0.28 / 1M |
Blended Price Weighted 7:2:1 cache:input:output per 1M | $0.14 / 1M | $0.07 / 1M |
Context Window Maximum input + output tokens | 256K tokens~384 A4 pages of text | 128K tokens~192 A4 pages of text |
Max Output Tokens Per single request | 8K | 8K |
Throughput Speed Tokens generated per second | 110 tok/s | 74 tok/s |
SWE-bench Verified Real-world software engineering | 28% | 42.1% |
GPQA Diamond Graduate-level reasoning | 42% | 49.5% |
Cached Input Price Prompt caching discount per 1M tokens | $0.05 / 1M | $0.01 / 1M |
Chi vince in ogni categoria?
At-a-glance decision frameworkPunteggi Benchmark Verificati
Official evaluation metrics verified by independent third-party research labs.
Industry benchmark evaluating real-world GitHub issue resolution and multi-file debugging.
Challenging questions created by domain experts to test complex PhD-level reasoning.
Hardened reasoning and academic knowledge benchmark with reduced lucky-guess probability.
Crowdsourced human preference evaluation based on blind, head-to-head model battles.
Extremely difficult, closed-ended questions spanning the frontier of human expert knowledge.
Autonomous multi-step tasks executed inside a real command-line terminal environment.
Research-grade coding problems drawn from real scientific and numerical workloads.
Contamination-free competitive programming problems collected continuously over time.
American Invitational Mathematics Examination — multi-step Olympiad-level reasoning.
Accuracy when reasoning over very large inputs that fill the model context window.
Tokens emitted per second during active generation streaming.
Latenza e Tempo di Risposta
Time to first token (TTFT) and estimated end-to-end response time for a typical 500-token response.
Simulatore Interattivo dei Costi Token
Adjust sliders to calculate projected monthly cloud API spending with prompt caching discounts.
Codestral · Strengths & Key Advantages
- Finestra di contesto da 256K con capacità massima di output fino a 8K token (SWE-bench Verified 28% · GPQA Diamond 42%)
- Throughput elevato di 110 token/secondo con latenza al primo token (TTFT) di soli 350 ms
- Pesi aperti (22B) pronti per il self-hosting locale o su cloud privato tramite vLLM, SGLang o Ollama
Considerations & Trade-offs
- Richiede server GPU dedicati con sufficiente memoria VRAM per l’esecuzione locale a piena precisione
- Nelle dimostrazioni scientifiche più estreme, i modelli di ragionamento pesante dedicati offrono maggiore profondità
DeepSeek-Coder-V2 236B · Strengths & Key Advantages
- Finestra di contesto da 128K con capacità massima di output fino a 8K token (SWE-bench Verified 42.1% · GPQA Diamond 49.5%)
- Throughput elevato di 74 token/secondo con latenza al primo token (TTFT) di soli 510 ms
- Pesi aperti (236B) pronti per il self-hosting locale o su cloud privato tramite vLLM, SGLang o Ollama
Considerations & Trade-offs
- Richiede server GPU dedicati con sufficiente memoria VRAM per l’esecuzione locale a piena precisione
- Nelle dimostrazioni scientifiche più estreme, i modelli di ragionamento pesante dedicati offrono maggiore profondità
Token Use & Task Economics
Estimated cost and token consumption to complete one representative agentic task.
| Metric | Codestral | DeepSeek-Coder-V2 236B |
|---|---|---|
Cost per Task Estimated USD to finish one task | $0.20 | $0.14 |
Output Tokens / Task Tokens generated per task | 24K | 30K |
Reasoning Tokens / Task Thinking tokens burned per task | 0K | 0K |
Specifiche Tecniche
Architecture, release timeline, capabilities, and modalities.
| Feature | Codestral | DeepSeek-Coder-V2 236B |
|---|---|---|
| Official API Model ID (SDK) | codestral-latest | deepseek-chat |
| API Protocol | OpenAI-Compatible REST / vLLM / SGLang / HuggingFace TGI | OpenAI-Compatible REST / vLLM / SGLang / HuggingFace TGI |
| Release Date | 2024-05 | 2024-06 |
| Knowledge Cutoff | May 2024 | December 2023 |
| Architecture | Code-Optimized Transformer with 32K Window | 236B Mixture-of-Experts (21B active) |
| Parameters | 22B | 236B |
| Enterprise Rate Limits (RPM / TPM) | Unlimited (Hardware-bound on self-hosted GPU) / 10,000 RPM Cloud · Unlimited local throughput / 10M+ TPM Managed Cloud | Unlimited (Hardware-bound on self-hosted GPU) / 10,000 RPM Cloud · Unlimited local throughput / 10M+ TPM Managed Cloud |
| Long-Context Tier (>128K/200K) | Flat rate ($0.2/1M in) across full 256K window | Flat rate ($0.14/1M in) across full 128K window |
| Prompt Cache Policy & TTL | Cached input at $0.05/1M tokens | Automatic prefix caching at $0.014/1M (no extra cache-write fee) |
| Computer Use / GUI Agent | API / CLI Tool-Use & DOM Agent ready (no raw pixel desktop click loop) | API / CLI Tool-Use & DOM Agent ready (no raw pixel desktop click loop) |
| MCP & Parallel Tool Calling | Native MCP Client/Server ready · Parallel function calling · 100% Strict JSON Schema | Native MCP Client/Server ready · Parallel function calling · 100% Strict JSON Schema |
| Web Search Grounding & Sandbox | Supported via external Tavily/Exa Search & E2B/Docker Sandbox tools | Supported via external Tavily/Exa Search & E2B/Docker Sandbox tools |
| Official Fine-Tuning Support | Full Weights SFT, LoRA, QLoRA, DPO & GRPO RL (Unsloth / Axolotl / Llama-Factory) | Full Weights SFT, LoRA, QLoRA, DPO & GRPO RL (Unsloth / Axolotl / Llama-Factory) |
| Multimodal Token Billing | Text BPE (~4 chars/token) | Text BPE (~4 chars/token) |
| Quantization & Min GPU VRAM | BF16 / FP16, FP8 Dynamic, INT4 AWQ / GPTQ, GGUF (Q4_K_M / Q8_0) · 14 GB VRAM (4-bit Quantized) · 47 GB VRAM (Native BF16) | BF16 / FP16, FP8 Dynamic, INT4 AWQ / GPTQ, GGUF (Q4_K_M / Q8_0) · 146 GB VRAM (4-bit Quantized) · 507 GB VRAM (Native BF16) |
| Safety & Frontier Classification | Open-Weights System Card + Customizable Guardrail Layer (Llama Guard / ShieldGemma) | Open-Weights System Card + Customizable Guardrail Layer (Llama Guard / ShieldGemma) |
| Enterprise Compliance (ZDR/HIPAA/SOC2) | 100% Air-Gapped On-Premises / Private VPC · Full HIPAA, GDPR & SOC2 Control | 100% Air-Gapped On-Premises / Private VPC · Full HIPAA, GDPR & SOC2 Control |
| Input Modalities | text | text |
| Output Modalities | text | text |
Codestral vs DeepSeek-Coder-V2 236B: Risposte Rapide (FAQ)
Which is more intelligent, Codestral or DeepSeek-Coder-V2 236B?
DeepSeek-Coder-V2 236B is more intelligent. It scores 42 compared with 40 on the composite Intelligence Index.
Which model is faster?
Codestral is faster, generating 110 tokens per second compared with 74 tokens per second.
Which model is cheaper to run?
DeepSeek-Coder-V2 236B is cheaper at $0.07 per 1M tokens blended, compared with $0.14 (7:2:1 cache:input:output ratio).
Which model has lower latency?
Codestral has lower latency, with a time to first token of 0.35s compared with 0.51s.
Which model has a larger context window?
Codestral has a larger context window, supporting 256K tokens compared with 128K tokens.
Verdetto
Sia Codestral (Ideale per lo sviluppo software e la generazione di codice con pesi aperti auto-ospitabili, sistemi RAG aziendali e integrazioni API scalabili.) sia DeepSeek-Coder-V2 236B (Ideale per il self-hosting su infrastruttura privata, il fine-tuning e la piena sovranità dei dati, sistemi RAG aziendali e integrazioni API scalabili.) eccellono in produzione: DeepSeek-Coder-V2 236B guida nei benchmark di ragionamento e programmazione, Codestral garantisce una generazione più rapida e DeepSeek-Coder-V2 236B offre il miglior rapporto qualità-prezzo per carichi API elevati.
Domande Frequenti (FAQ)
Quale modello è migliore per coding e ragionamento, Codestral o DeepSeek-Coder-V2 236B?+
Nell’indice di intelligenza composito e nei benchmark di programmazione prevale DeepSeek-Coder-V2 236B. Consulta i punteggi SWE-bench e GPQA qui sopra per tutti i dettagli.
Qual è la differenza di prezzo API tra Codestral e DeepSeek-Coder-V2 236B?+
Codestral costa $0.20 / $0.60 per milione di token di input/output, mentre DeepSeek-Coder-V2 236B costa $0.14 / $0.28. Nel costo combinato, DeepSeek-Coder-V2 236B risulta più economico.
Quale modello ha la finestra di contesto più ampia e la velocità maggiore?+
Codestral supporta 256K token di contesto contro 128K token di DeepSeek-Coder-V2 236B, mentre Codestral offre una velocità di generazione superiore.
Understand key metrics like context window, token pricing, reasoning mode, and SWE-bench scores in our comprehensive AI glossary.
Related reading
- Devstral Small vs Codestral: Confronto Benchmark, Prezzi API e Velocità (2026)Confronta
- Qwen 2.5 Coder 32B vs DeepSeek-Coder-V2 236B: Confronto Benchmark, Prezzi API e Velocità (2026)Confronta
- Aion-3.0 vs Aion-Mini: Confronto Benchmark, Prezzi API e Velocità (2026)Confronta
- Aion-3.0 vs Gemini 3.6 Flash: Confronto Benchmark, Prezzi API e Velocità (2026)Confronta
- Aion-Mini vs Laguna XS 2.1: Confronto Benchmark, Prezzi API e Velocità (2026)Confronta
- Aya Expanse 32B vs Llama 3.2 90B Vision: Confronto Benchmark, Prezzi API e Velocità (2026)Confronta
Benchmark scores, token prices, and rate limits evolve rapidly as AI labs release model updates. Always verify current specifications and pricing on the official provider documentation.
Fonti ufficiali: Documentazione API ufficiale di Mistral AI e DeepSeek, schede di sistema e benchmark indipendenti (2026).