PERFORMANCELlama 3.3 70B → Scaleway: throughput ↑ 53%6hPERFORMANCEGLM-5.2 → Scaleway: TTFT ↓ 30%6hPERFORMANCEMistral Medium 3.5 → Scaleway: throughput ↑ 22%6hPERFORMANCEQwen3-235B → Scaleway: throughput ↑ 67%6hPERFORMANCEGPT-5.6 Sol → OpenAI: TTFT ↓ 22%6hPERFORMANCEMistral Large 3 → Mistral AI: TTFT ↓ 21%6hPERFORMANCEGPT-5.6 Luna → OpenAI: throughput ↑ 21%6hPERFORMANCEDeepSeek V4 Flash → Scaleway: throughput ↑ 147%6hPERFORMANCEDeepSeek V4 Flash → Scaleway: TTFT ↓ 71%6hPERFORMANCEGPT-OSS 120B → Scaleway: TTFT ↓ 46%6hPERFORMANCEGPT-OSS 120B → Together AI: TTFT ↓ 40%6hPERFORMANCEKimi K3 → Together AI: TTFT ↓ 20%6hPERFORMANCEDeepSeek V4 Flash → Scaleway: throughput ↑ 36%30hPERFORMANCEDeepSeek V4 Flash → Scaleway: TTFT ↓ 19%30hPERFORMANCEQwen3-235B → Scaleway: throughput ↑ 19%2dPERFORMANCEDeepSeek V4 Flash → Scaleway: throughput ↑ 58%2dPERFORMANCEDeepSeek V4 Flash → Scaleway: TTFT ↑ 21%2dPERFORMANCEGPT-5.6 Sol → OpenAI: throughput ↑ 16%2dPERFORMANCEGPT-OSS 120B → Together AI: TTFT ↑ 22%2dPERFORMANCEKimi K3 → Together AI: throughput ↑ 62%2dPERFORMANCEKimi K3 → Together AI: TTFT ↓ 42%2dPERFORMANCEQwen3-235B → Scaleway: throughput ↑ 30%3dPERFORMANCEDeepSeek V4 Flash → Scaleway: TTFT ↓ 56%3dPERFORMANCELlama 3.3 70B → Scaleway: throughput ↑ 34%3dPERFORMANCELlama 3.3 70B → Scaleway: TTFT ↓ 26%3dPERFORMANCEKimi K3 → Together AI: throughput ↑ 128%3dPERFORMANCEGPT-OSS 120B → Together AI: throughput ↑ 29%3dPERFORMANCEGPT-OSS 120B → Together AI: TTFT ↓ 21%3dPERFORMANCEGPT-OSS 20B → Groq: TTFT ↑ 66%3dPERFORMANCEGLM-5.2 → Together AI: throughput ↑ 47%3dPERFORMANCEMistral Medium 3.5 → Scaleway: TTFT ↑ 18%4dPERFORMANCEGLM-5.2 → Scaleway: throughput ↓ 27%4dPERFORMANCEGLM-5.2 → Scaleway: TTFT ↑ 49%4dPERFORMANCEDeepSeek V4 Flash → Scaleway: throughput ↓ 71%4dPERFORMANCEDeepSeek V4 Flash → Scaleway: TTFT ↑ 190%4dPERFORMANCEGPT-5.6 Luna → OpenAI: throughput ↓ 23%4dPERFORMANCELlama 3.3 70B → Scaleway: TTFT ↑ 60%4dPERFORMANCEGLM-5.2 → Together AI: TTFT ↓ 25%4dPERFORMANCEGPT-OSS 120B → Together AI: throughput ↓ 16%4dPERFORMANCEGPT-OSS 120B → Together AI: TTFT ↓ 17%4d
ENGINEERING STUDY

Structured Output: Trust, but Validate

Schema-conformance rates across measured providers, one mechanical validator, no partial credit.

Key findings
  • Measured structured-output validity spans 91.7%–100.0% across providers.
  • Cerebras leads; Mistral trails — a pipeline parsing its output re-requests or crashes on the difference.

Conformance, measured

Each structured-output case demands schema-valid JSON and validates the result mechanically — no partial credit. The native-schema column records each provider's declared enforcement mode; whether it separates the table is for the measured column to say. Anthropic runs this suite prompt-only by recorded request quirk.

ProviderValidityNative schema mode
Cerebras100.0%no
Groq100.0%yes
Together100.0%yes
OpenAI100.0%yes
Scaleway100.0%yes
Mistral91.7%yes
Structured-output validity by measured provider · ● MEASURED
What this means for engineers
  • Treat schema validity as an SLO with a measured baseline per provider — and always validate downstream regardless.
Dataset

Figures recompute from the live rolling window on every visit; the snapshot and versions above are the provenance of the regime that produced them. View the current benchmark →

Cite this researchInferenceBench (2026). Structured Output: Trust, but Validate. InferenceBench Research.
← All research
Related research
Structured Output: Trust, but Validate · InferenceBench