Terminal-Bench 4.0 — Mercor evaluation (Terminal-Bench 4.0 (Mercor))

We show this table for reference; we do not rank on it.

Completing command-line tasks on the Terminal-Bench 4.0 set. This table shows Mercor-run configurations for reference and is excluded from model rankings.

Evaluation results published by Mercor. Published evaluation aggregates only; task contents and dataset license grants are not included.

Pass@1 on Terminal-Bench 4.0 (Mercor) — October 1, 2026 capture

We mirror the published pass@1 view for Terminal-Bench 4.0 (Mercor). Opus 5.5 leads the public snapshot at 58.30%, followed by GPT-6.1 Sol (58.30%) and GPT-6 Astra (54.50%). We do not use these results to rank models overall.

Opus 5.5

Anthropic

mini-swe-agent (1000 max steps, 8 hour time limit) · max reasoning

Published margin ±10.9 · 192 reported samples

58.30%

GPT-6.1 Sol

OpenAI

mini-swe-agent (1000 max steps, 8 hour time limit) · max reasoning

Published margin ±10.5 · 196 reported samples

58.30%

GPT-6 Astra

OpenAI

mini-swe-agent (1000 max steps, 8 hour time limit) · max reasoning

Published margin ±10.1 · 198 reported samples

54.50%

21 configurationsAgenticDated Mercor configurationsDisplay onlyUpdated October 1, 2026 capture

Pass@1 table (21 configurations)

Score

Opus 5.5Anthropicmini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.9 · 192 reported samples

58.30%

GPT-6.1 SolOpenAImini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.5 · 196 reported samples

58.30%

GPT-6 AstraOpenAImini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.1 · 198 reported samples

54.50%

Fable 5.1Anthropicmini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.6 · 198 reported samples

53.00%

Sonnet 5.5Anthropicmini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.9 · 174 reported samples

52.50%

Opus 5Anthropicmini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.1 · 198 reported samples

49.50%

GPT-6 SolOpenAImini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.4 · 191 reported samples

43.80%

GLM-5.3Zhipumini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±10.1 · 198 reported samples

37.90%

GPT-5.6 Sol (Pro)OpenAImini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±8.8 · 198 reported samples

32.30%

GLM-5.3-FlashZhipumini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±8.8 · 197 reported samples

27.30%

GPT-5.6 TerraOpenAImini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±8.3 · 198 reported samples

27.30%

Kimi K3Kimimini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±8.6 · 198 reported samples

26.30%

Grok 4.6xAImini-swe-agent (1000 max steps, 8 hour time limit) · xhigh reasoningPublished margin ±8.3 · 198 reported samples

25.30%

GPT-6 LunaOpenAImini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±7 · 192 reported samples

13.50%

DeepSeek-V4-Pro-0813DeepSeekmini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±5.3 · 198 reported samples

12.60%

Gemini 3.8 FlashGooglemini-swe-agent (1000 max steps, 8 hour time limit) · high reasoningPublished margin ±7.1 · 198 reported samples

12.10%

GPT-5.6 LunaOpenAImini-swe-agent (1000 max steps, 8 hour time limit) · max reasoningPublished margin ±5.8 · 198 reported samples

11.60%

Grok 4.7xAImini-swe-agent (1000 max steps, 8 hour time limit) · xhigh reasoningPublished margin ±4.5 · 198 reported samples

6.60%

MiniMax-M3MiniMaxmini-swe-agent (1000 max steps, 8 hour time limit) · high reasoningPublished margin ±1.3 · 197 reported samples

1.00%

Gemini 3.5 Flash LiteGooglemini-swe-agent (1000 max steps, 8 hour time limit) · high reasoningPublished margin ±0.8 · 198 reported samples

0.50%

InklingThinking Machinesmini-swe-agent (1000 max steps, 8 hour time limit) · high reasoningPublished margin ±0.8 · 198 reported samples

0.50%

How to read this Mercor evaluation

We captured Mercor's Terminal-Bench 4.0 results on October 1, 2026. The table preserves 21 published configurations, their source identifiers, reasoning settings, and disclosed pass@1 values. These are results from Mercor's evaluation setup; we did not rerun them.

This dated source table is display only and does not enter overall or category rankings. Source configurations retain their external identity, including Pro descriptors and different effort settings.

The headline is the source page's selected Pass@1 metric. The task set, harness, tools, and grader belong to this Mercor run and may differ from another evaluator's results.

The row labels retain published error margins and sample counts when supplied. Missing margins and counts remain unreported; a sample count is not used to reconstruct a task denominator. Except where the dedicated source defines an interval, the confidence level is unreported. Equal displayed scores and overlapping margins do not establish statistical ties. Capture time does not establish individual evaluation dates.

The general methodology lists programmatic unit tests (all must pass) and a judge of None. Its run-count label is k = 3. The linked result page takes precedence where its setup differs, as noted above.

Snapshot

Pass@121 configurationsMercor evaluationDisplay only

Mercor Terminal-Bench 4.0 results Mercor APEX benchmark index Mercor methodology Source paper Upstream evaluation code Upstream dataset

The published Terminal-Bench 4.0 (Mercor) snapshot places Opus 5.5 first at 58.30%. The third row is 3.80 points behind. The broader top-10 range is 31.00 points, so the table still separates the published systems.

21 configurations are shown for Terminal-Bench 4.0 (Mercor). The benchmark falls in the Agentic category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. Terminal-Bench 4.0 (Mercor) is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About Terminal-Bench 4.0 (Mercor)

Tasks

66 source-reported tasks

Format

Pass@1

Difficulty

Source-specific evaluation

The selected headline is Pass@1. Source model identifiers, effort settings, published margins, and reported sample counts remain attached to the source configuration. Public-set results and Mercor task extensions occupy separate tables.

Terminal-Bench 4.0 evaluation results Public benchmark source

Freshness and provenance

Version

Terminal-Bench 4.0; Mercor Pass@1

Refresh cadence

Reviewed dated source capture

Staleness state

Dated Mercor configurations

Question availability

Upstream task access and terms; only aggregate results mirrored

Dated Mercor configurationsDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Questions

What does Terminal-Bench 4.0 (Mercor) measure?

Completing command-line tasks on the Terminal-Bench 4.0 set. This table shows Mercor-run configurations for reference and is excluded from model rankings.

Which model leads the published Terminal-Bench 4.0 (Mercor) snapshot?

Opus 5.5 currently leads the published Terminal-Bench 4.0 (Mercor) snapshot with 58.30% pass@1. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on Terminal-Bench 4.0 (Mercor)?

The October 1, 2026 capture snapshot contains 21 source configurations.

Last updated: October 1, 2026 capture · mirrored from the public benchmark leaderboard

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.