Capacité · 2026-06-29

Modèles d'IA avec support de la sortie structurée

Modèles avec JSON mode ou sortie conforme à un schéma pour des pipelines d'extraction et de classification fiables.

Qu'est-ce que c'est ?

La sortie structurée (JSON mode ou response_format=json_schema) contraint le modèle à émettre un document JSON conforme à votre schéma.
Contrairement à un prompt demandant « réponds en JSON », c'est appliqué au décodage — le modèle ne peut pas émettre de JSON invalide.

Pourquoi c'est important

Réduit les erreurs de parsing et les préfixes comme « Bien sûr, voici le JSON… ».
Indispensable pour les pipelines qui alimentent des systèmes typés : extraction, classification, résumés structurés.

306 modèles avec cette capacité

Modèle	Éditeur	Entrée / 1M	Sortie / 1M	Contexte	Fournisseurs
Voxtral Small 24B 2507	Mistral	$0.002	$0.002	32K	4
Ling-2.6-flash	openrouter	$0.010	$0.030	262K	1
Gemma 3n 4B	Google	$0.020	$0.040	33K	4
Llama 3 8B Lunaris	Meta	$0.040	$0.050	8K	2
Sao10k L3 8B Lunaris	novita-ai	$0.050	$0.050	8K	1
MythoMax 13B	openrouter	$0.060	$0.060	4K	1
Google Gemma 3 27B Instruct	Google	$0.030	$0.110	203K	10
Google Gemma 3 12B	Google	$0.050	$0.100	131K	7
Granite 4.1 8B	openrouter	$0.050	$0.100	131K	1
Granite 4.1 8B	nano-gpt	$0.050	$0.100	131K	1
gpt-oss-20b	OpenAI	$0.029	$0.140	128K	24
Qwen3 235B A22B 2507	Alibaba (Qwen)	$0.071	$0.100	262K	3
gpt-oss-120b	OpenAI	$0.030	$0.150	128K	37
Qwen3.5 9B	Alibaba (Qwen)	$0.040	$0.150	262K	14
GPT OSS 20B	llmgateway	$0.040	$0.150	131K	1
Trinity Mini	openrouter	$0.045	$0.150	131K	1
Ministral 3 3B 2512	Mistral	$0.100	$0.100	131K	3
Ministral 3B	llmgateway	$0.100	$0.100	131K	1
Reka Edge	openrouter	$0.100	$0.100	16K	1
Mistral Small 3.2 24B	Mistral	$0.060	$0.180	128K	3
GPT OSS 20B	databricks	$0.050	$0.200	131K	1
GPT OSS 20B	neon	$0.050	$0.200	131K	1
GPT OSS Safeguard 20B	OpenAI	$0.070	$0.200	128K	6
Qwen2.5 VL 32B Instruct	Alibaba (Qwen)	$0.050	$0.220	131K	3
GPT OSS 20B	frogbot	$0.070	$0.200	131K	1
Hermes 2 Pro Llama 3 8B	Meta	$0.140	$0.140	8K	3
Qwen2.5 72B Instruct	Alibaba (Qwen)	$0.062	$0.231	33K	5
Ministral 3 8B 2512	Mistral	$0.150	$0.150	262K	3
Ministral 8B	llmgateway	$0.150	$0.150	262K	1
GPT OSS 120B	llmgateway	$0.050	$0.250	131K	1
Reka Flash 3	openrouter	$0.100	$0.200	66K	1
Ling 2.6 Flash	nano-gpt	$0.080	$0.240	262K	1
GPT OSS 120B	databricks	$0.072	$0.280	131K	1
GPT OSS 120B	neon	$0.072	$0.280	131K	1
Qwen3 30B A3B	Alibaba (Qwen)	$0.080	$0.280	41K	10
Seed 1.6 Flash (250715)	llmgateway	$0.070	$0.300	256K	1
Gemini 2.0 Flash-Lite	Google	$0.075	$0.300	1.05M	4
Seed 1.6 Flash	openrouter	$0.075	$0.300	262K	1
Llama 4 Scout	Meta	$0.080	$0.300	328K	5
Gemma 4 26B A4B IT	Google	$0.060	$0.330	262K	16
Gemma 4 31B IT	Google	$0.100	$0.300	262K	26
Ministral 3 14B 2512	Mistral	$0.200	$0.200	262K	3
Ling-2.6-flash	novita-ai	$0.100	$0.300	262K	1
XiaomiMiMo/MiMo-V2-Flash	novita-ai	$0.100	$0.300	262K	1
Ministral 14B	llmgateway	$0.200	$0.200	262K	1
Llama 3.2 11B Instruct	Meta	$0.070	$0.330	128K	1
OWL	nano-gpt	$0.100	$0.300	1.05M	1
Mistral Small 3.2 24B Instruct 2506	Mistral	$0.100	$0.310	32K	5
DeepSeek V4 Flash	DeepSeek	$0.140	$0.280	1M	31
Coding Router Low	nano-gpt	$0.140	$0.280	1M	1
Coding Router Medium	nano-gpt	$0.140	$0.280	1M	1
Phi 4 Mini Instruct	Microsoft	$0.080	$0.350	128K	5
GPT-4o-mini Search Preview	OpenAI	$0.088	$0.350	128K	5
Gemini 2.5 Flash Lite Preview 09-2025	Google	$0.090	$0.360	1.05M	6
Qwen3.5 Flash	Alibaba (Qwen)	$0.090	$0.360	1M	4
GPT-5 Nano	OpenAI	$0.050	$0.400	400K	22
Gemini 2.5 Flash-Lite	Google	$0.100	$0.400	1.05M	17
GPT-4.1 nano	OpenAI	$0.100	$0.400	1.05M	16
Gemini Flash-Lite Latest	Google	$0.100	$0.400	1.05M	5
Gemini 2.0 Flash	Google	$0.100	$0.400	1.05M	3

Top 60 sur 306 affichés. Utilisez le répertoire complet pour filtrer davantage.

Frequently asked questions

How many AI models support sortie structurée?

306 canonical models in our database currently support sortie structurée. The list is regenerated on every data refresh, so it always reflects the latest releases tracked in our catalogue.

What is the cheapest model with sortie structurée?

Voxtral Small 24B 2507 from Mistral is currently the lowest-priced option, at $0.002 per 1M input tokens and $0.002 per 1M output tokens. The full table above is sorted price-ascending.

Which model with sortie structurée has the largest context window?

Grok 4.20 Multi-Agent (xAI) leads on context at 2M tokens. This may matter if you also need long-document understanding alongside sortie structurée.

Which models are available on the most providers?

Production-readiness usually correlates with how many independent providers host the same weights. The top three by provider count are: Kimi K2.6 (49), Kimi K2.5 (48), GLM-5.1 (47).

How is sortie structurée different from a regular LLM?

Structured output (a.k.a. JSON mode / response_format=json_schema) constrains the model at decode time so it cannot emit invalid JSON. This is stricter than just prompting 'reply in JSON' and removes a whole class of parsing errors.

How often is this list updated?

Daily. Our data pipeline syncs once a day, regenerates the canonical model list, and rebuilds these pages so newly released models appear within 24 hours.

Top models with this capability

Voxtral Small 24B 2507$0.00 in / $0.00 out
Ling-2.6-flash$0.01 in / $0.03 out
Gemma 3n 4B$0.02 in / $0.04 out
Llama 3 8B Lunaris$0.04 in / $0.05 out
Sao10k L3 8B Lunaris $0.05 in / $0.05 out

Other capabilities

Best-of lists you might also want

Pricing comparisons

Vendors in this list

Dernière mise à jour : 2026-06-29

Prices in USD per 1M tokens. Unknown means the provider does not publish per-token pricing.

Pricing and capabilities are refreshed daily and reconciled against each provider's official documentation. Always verify critical production decisions with the provider directly.