LLM Reference
activeCodingAgents

ProgramBench Almost@1

Metric: Almost@1 (higher is better)Introduced: 2026

DeepSeek-reported ProgramBench evaluation from the DeepSeek V4.1 tech report Table 3, measuring Almost@1 with the DeepSeek Harness Minimal mode and a 1M-token context window. Keep separate from Moonshot's proprietary Program Bench benchmark used in Kimi launch materials.

Models ranked

1

tracked on this benchmark

Score band

20.3 – 20.3

best → lowest tracked

Snapshot trend

need ≥2 snapshots

Leaderboard

Tracked models ranked by Almost@1 (higher is better).

Compare candidates
#Model variant and provenanceScore
1
DeepSeek V4.1 Flash
Version: ProgramBench Almost@1; DeepSeek Harness Minimal mode; 1M contextHarness: Not recordedEvaluator: Not recordedObserved: Sep 10, 2026Confidence: Not recordedSource
20.3

How to read this benchmark

This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.

Trust this score when

  • There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
  • The model list covers the same version family you can actually deploy today.
  • Top candidates overlap with your required routing and feature requirements.

Be cautious when

  • There is only one benchmark snapshot or the dataset appears stale.
  • The benchmark metric direction is opposite of your decision objective.
  • The score difference between options is narrow and likely within implementation variance.

Related benchmarks

Last reviewed: Sep 10, 2026

Resources