Skip to content
Hbench
hbench

humaneval-plus@1.0 (unreviewed) on GPT-6 Astra

164 tasks, 1 attempt, open network. Updated 24 Oct 2026.

164 tasks from humaneval-plus@1.0 on ghcr.io. Nobody has reviewed this dataset, so a score here says nothing about the benchmark it's named after.

Build a run for this board

Value

What a task cost each configuration. Codex used 4% more tokens than Claude Code, and 10% less time per task.

$ per solved task

Lower is cheaper. Dollars per task divided by the pass rate.

Embed this chart

Harness$ per solved task
Claude Code2.1.309 · Reference$0.07$0.06 a task
Codex0.164.0$0.11$0.11 a task

Tokens per task

Lower is leaner. Input plus output, cached input included.

Embed this chart

HarnessTokens per task
Claude Code2.1.309 · Reference40k×1.00
Codex0.164.042k×1.04

Agent time per task

Lower is faster. Agent start to finish; setup and tests left out.

Embed this chart

HarnessTime per task
Codex0.164.021 s×0.90
Claude Code2.1.309 · Reference23 s×1.00

Pass rate against tokens per task

Up and to the left is better. Pass-rate axis starts at 90%.

Embed this chart

90%95%100%39k40k41k42kTokens per task, input and outputMore tokens for a lower pass rate than Claude CodeClaude Code2.1.309 · 40k · 96.3%Codex0.164.0 · 42k · 95.7%
  • On the frontier: no other row has a higher pass rate for less tokens
  • Off the frontier
  • 95% interval of the pass rate
On the frontier, least tokens first: Claude Code 2.1.309, 40k tokens a task, 96.3%.
  • 2 runs
  • Updated 24 Oct 2026
  • Method