GPT-5.3 Codex
| Released | 2026-02-05 |
|---|---|
| Price per 1M | $1.75 in · $14.00 out |
| Overall score | 58.3 · rank 48 |
| Context window | 400,000 tokens |
| Provider | OpenAI |
| Weights | Proprietary |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 58.3 | 48 | 9 |
| Overall: Coding | 57.4 | 54 | 9 |
| Overall: Reasoning | 61.6 | 23 | 1 |
| Overall: Agentic work | 58.6 | 49 | 7 |
| Overall: Frontend | 55.0 | 61 | 5 |
| Coding | 57.2 | 56 | 3 |
| Agents | 58.4 | 48 | 3 |
| Reasoning | 61.6 | 32 | 1 |
| Frontend | 53.9 | 59 | 2 |
| Value vs price | +5.9 | 66 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 55.5 ±5.7 | 4 |
| High | 54.9 ±6.2 | 1 |
| Extra High | 58.4 ±4.2 | 5 |
Evidence
Coding 57.2 ±2.9
| Benchmark | Result | Source | Date |
|---|---|---|---|
| IOI | 53.8% | Vals AI | 2026-10-29 |
| LiveCodeBench (Vals run) | 87.3% | Vals AI | 2026-09-01 |
| SWE-bench Verified (Epoch run) | 74.8% | Epoch AI | 2026-02-25 |
| SWE-bench Verified (Vals run) | 78.0% | Vals AI | 2026-09-01 |
Agents 58.4 ±4.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench 2.0 | 78.4% | Epoch AI | 2026-02-05 |
| Vending-Bench 2 | $5,940 | Epoch AI | 2026-11-11 |
| METR 50% Time Horizon | $350 | METR | 2026-11-11 |
| METR task suite | 74.5% | Epoch AI | 2026-11-11 |
Reasoning 61.6 ±5.3
Frontend 53.9 ±0.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Vibe Code Bench v1.1 | 61.8% | Vals AI | 2026-11-09 |
| LMArena WebDev | 1,406 | LMArena | 2026-11-09 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| Terminal-Bench 2.0 | 77.3% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 75.8% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 75.7% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 75.1% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 74.6% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 71.5% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 70.3% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 69.1% | Epoch AI | A stronger configuration result was used |
| Terminal-Bench 2.0 | 64.7% | Epoch AI | A stronger configuration result was used |
Names matched to this model
gpt-5-3-codex