OpenAI/GPT-5/Mar 3, 2026
GPT-5.3 Instant
The conversational layer of GPT-5.3, now the default model in ChatGPT. Focuses on UX quality — 26.8% fewer hallucinations on web queries, fewer unnecessary refusals, and more natural writing. Supports reasoning effort levels from minimal (5 tokens) to high (200-700 tokens).
textcodereasoningvisiontool-useaudio
Arena ELO
—
Input Price
—
Output Price
—
Speed
—
Context
256K
Latency
—
Capability Assessment
No benchmark data available.
Comparative Analysis
| Metric | GPT-5.3 Instant | Claude Opus 4.6 | Gemini 3.1 Pro | Grok 4 |
|---|---|---|---|---|
| SWE-bench | — | 80.8% | 80.6% | 72.0% |
| AIME 2025 | — | 100.0% | 91.2% | 94.0% |
| GPQA Diamond | — | 91.3% | 94.3% | 88.0% |
| MMLU | — | 91.0% | 92.6% | — |
| Input $/M | — | $5 | $2 | $3 |
| Output $/M | — | $25 | $12 | $15 |