Loading benchmark catalogue
Compare performance benchmarks, coding intelligence, reasoning evaluations, context limits, and API economics between GPT-6 Astra (OpenAI) and Step 5 Preview (StepFun).
| Metric / Capability | GPT-6 Astra | Step 5 Preview |
|---|---|---|
| Developer / Organization | OpenAI | StepFun |
| Release Date | 2026-09-04 | 2026-10-08 |
| Deployment / Weights | Proprietary Cloud API | Open Weights |
| Context Window | 1.1M tokens (1,050,000 tokens) | 1.0M tokens (1,048,576 tokens) |
| Input Pricing (/1M Tokens) | $10 | Free (Open Weights) |
| Output Pricing (/1M Tokens) | $50 | Free (Open Weights) |
| Benchmark / Domain | GPT-6 Astra | Step 5 Preview | Advantage |
|---|---|---|---|
Artificial Analysis Intelligence Index Composite index aggregating 10+ evaluations of general reasoning and knowledge. | 52.7% | 43.7% | GPT-6 Astra (+21%) |
Artificial Analysis Coding Index Composite index of coding and software-engineering evaluations. | 76.9% | — | — |
Artificial Analysis Agentic Index Composite index of tool-use and long-horizon agentic tasks. | 51.0% | — | — |
Artificial Analysis Coding Agent Index v2.0 Composite index of coding-agent systems (model + harness) across FrontierCode v1.1, SWE-Atlas-QnA, and Terminal-Bench v4. | 54.7% | 43.1% | GPT-6 Astra (+27%) |
Terminal-Bench 4.0 Sandboxed command-line tasks requiring shell and file-system operation. | 57.9% | 33.3% | GPT-6 Astra (+74%) |
ARC-AGI-2 Novel visual grid puzzles requiring abstract rule induction. | 95.0% | — | — |
GPQA Diamond Graduate-level physics, chemistry, and biology questions written by domain experts. | 96.0% | 93.5% | GPT-6 Astra (+3%) |
Humanity's Last Exam Expert-written questions across hundreds of disciplines designed to resist saturation. | — | 46.5% | — |
FrontierMath v2 (Tiers 1-3) Unpublished, research-level mathematics problems across Tiers 1-3. | 93.7% | — | — |
GPT-6 Astra leads across 4 evaluated benchmarks compared to 0 for Step 5 Preview, showing notable advantages in Artificial Analysis Intelligence Index, Artificial Analysis Coding Agent Index v2.0, Terminal-Bench 4.0.
GPT-6 Astra offers a larger context capacity of 1.1M tokens (1,050,000 tokens), compared to 1.0M tokens for Step 5 Preview.
Step 5 Preview is more economical. Input tokens cost $Free (Open Weights)/M tokens compared to $10/M tokens for GPT-6 Astra.
GPT-6 Astra is a proprietary cloud API. Step 5 Preview is an open-weights model.
Analyze 130+ models side-by-side using interactive radar charts, cost-efficiency quadrant plots, and live dataset filters on the full NeuBench index.
Open Interactive Comparator