GPT-5.6 LunavsDeepSeek-V4-Flash
Across 6 shared benchmarks, GPT-5.6 Luna leads overall: GPT-5.6 Luna wins 5, DeepSeek-V4-Flash wins 1, with 0 ties and an average score difference of +3.97.
GPT-5.6 Luna
OpenAI · 2026-06-26 · Reasoning model
DeepSeek-V4-Flash
DeepSeek-AI · 2026-04-24 · Reasoning model
GPT-5.6 Luna5 wins(83%)(17%)1 winDeepSeek-V4-Flash
Benchmark scores
Grouped by capability, sorted by largest gap within each. 6 shared benchmarks.
Coding and Software Engineer
GPT-5.6 Luna 2/3| Benchmark | GPT-5.6 Luna | DeepSeek-V4-Flash | Diff |
|---|---|---|---|
| Text Arena (Coding) | 1,52315 / 35极高强度思考（无工具） | 1,5776 / 35Thinking High (No Tools) | -53.60 |
| WeirdML v2 | 60.8627 / 52Thinking High (With Tools) | 43.7644 / 52Thinking High (With Tools) | +17.10 |
| DeepSWE | 67.206 / 27极高强度思考（工具） | 54.4015 / 27Max (With Tools) | +12.80 |
Agent Level Benchmark
GPT-5.6 Luna 1/1| Benchmark | GPT-5.6 Luna | DeepSeek-V4-Flash | Diff |
|---|---|---|---|
| Agents' Last Exam | 50.303 / 11极高强度思考（工具） | 25.2010 / 11Max (With Tools) | +25.10 |
AI Agent - Tool Usage
GPT-5.6 Luna 1/1| Benchmark | GPT-5.6 Luna | DeepSeek-V4-Flash | Diff |
|---|---|---|---|
| Terminal-Bench 2.1 | 84.7010 / 44最高（无工具） | 82.7016 / 44Max (With Tools) | +2 |
General Evaluation
GPT-5.6 Luna 1/1| Benchmark | GPT-5.6 Luna | DeepSeek-V4-Flash | Diff |
|---|---|---|---|
| GPQA Diamond | 91.6027 / 226最高（无工具） | 71.20151 / 226Normal (No Tools) | +20.40 |
Specs
| Field | GPT-5.6 Luna | DeepSeek-V4-Flash |
|---|---|---|
| Publisher | OpenAI | DeepSeek-AI |
| Release date | 2026-06-26 | 2026-04-24 |
| Model type | Reasoning model | Reasoning model |
| Architecture | Dense | MoE |
| Parameters | Not available | 284B |
| Context length | 1.05M | 1M |
| Max output | 128K | 384K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | GPT-5.6 Luna | DeepSeek-V4-Flash |
|---|---|---|
| Text input | $1 / 1M tokens | $0.14 / 1M tokens |
| Text output | $6 / 1M tokens | $0.28 / 1M tokens |
| Cache read | $0.1 / 1M tokens | $0.0028 / 1M tokens |
| Cache write | $1.25 / 1M tokens | Not public |
Summary
- GPT-5.6 Lunaleads in:Coding and Software Engineer (2/3), Agent Level Benchmark (1/1), AI Agent - Tool Usage (1/1), General Evaluation (1/1)
On average across the 6 shared benchmarks, GPT-5.6 Luna scores 3.97 higher.
Largest single-benchmark gap: Text Arena (Coding) — GPT-5.6 Luna 1,523 vs DeepSeek-V4-Flash 1,577 (-53.60).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.