Benchmark Almanac

DPAI Arena Spring Benchmark

Claude Opus 1M (max ultrathink) 66.0 tops its leaderboard Claude Opus 4.7 64.9 leads the models we track
Not in model cards

Coding agents implement changes in Java Spring repositories. DPAI Arena scores each run and publishes its duration.

The leaderboard reports runs with at least five distinct models, including GPT-5.5, GPT-5.4, Gemini 3 Flash, Claude Opus 4.7 and Grok 4.1 Fast Reasoning.

Made by
JetBrains and Toloka
Released
October 2025
Kind
Fixing real repositories
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Code, Actions
Interaction
Agentic
Environment
Terminal
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
In views
Agents
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
15
Entries on its own leaderboard, 7 models we track

Claude Opus 1M (max ultrathink) tops the leaderboard with 66.0 Claude Opus 4.7 leads with 64.9 among models released by April 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. The best model we track is Claude Opus 4.7, at 64.9. Agents built around a model, such as Junie CLI* 1512.1 + Gemini 3.1 Pro, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

The newest model with a score here came out in April 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Claude Opus 1M (max ultrathink) max ultrathink 66.0
  2. 2 Claude Opus 4.7 max ultrathink 64.9
  3. 3 GPT-5.4 xHigh 64.7
  4. 4 GPT-5.5 xHigh 64.1
  5. 5 Junie CLI* 1512.1 + Gemini 3.1 Pro 63.7
  6. 6 Junie CLI* 1512.1 + Gemini 3 Flash 63.3
  7. 7 Gemini 3.1 Pro 62.6
  8. 8 Junie CLI* 1512.1 + Claude Opus 4.6 max 62.5
  9. 9 Claude Opus 4.6 max ultrathink 62.4
  10. 10 Junie CLI* 1512.1 + Claude Sonnet 4.6 max 62.2
  11. 11 Gemini 3 Flash 61.9
  12. 12 Junie CLI* 1512.1 + GPT-5.4 xhigh 60.8
  13. 13 Junie CLI* 1512.1 + GPT-5.3 Codex xhigh 60.6
  14. 14 Claude Sonnet 4.6 max ultrathink 60.1
  15. 15 Junie CLI* 1512.1 + Grok 4.1 Fast reasoning 47.7

Source: Official DPAI Arena Spring Benchmark leaderboard

#EntryScoreSettingAgent byDateLinkSource
1 Claude Opus 1M (max ultrathink) 66.0 max ultrathink, Claude Code 2.1.119 Anthropic dpaia.teamcity.com Official leaderboard
2 Claude Opus 4.7 64.9 max ultrathink, Claude Code 2.1.118 Anthropic Apr 2026 dpaia.teamcity.com Official leaderboard
3 GPT-5.4 64.7 xHigh, Codex CLI 0.124.0 OpenAI Mar 2026 dpaia.teamcity.com Official leaderboard
4 GPT-5.5 64.1 xHigh, Codex CLI 0.125.0 OpenAI Apr 2026 dpaia.teamcity.com Official leaderboard
5 Junie CLI* 1512.1 + Gemini 3.1 Pro 63.7 JetBrains Feb 2026 dpaia.teamcity.com Official leaderboard
6 Junie CLI* 1512.1 + Gemini 3 Flash 63.3 JetBrains Dec 2025 dpaia.teamcity.com Official leaderboard
7 Gemini 3.1 Pro 62.6 Gemini CLI 0.39.0 Google Feb 2026 dpaia.teamcity.com Official leaderboard
8 Junie CLI* 1512.1 + Claude Opus 4.6 62.5 max JetBrains Feb 2026 dpaia.teamcity.com Official leaderboard
9 Claude Opus 4.6 62.4 max ultrathink, Claude Code 2.1.119 Anthropic Feb 2026 dpaia.teamcity.com Official leaderboard
10 Junie CLI* 1512.1 + Claude Sonnet 4.6 62.2 max JetBrains Feb 2026 dpaia.teamcity.com Official leaderboard
11 Gemini 3 Flash 61.9 Gemini CLI 0.39.0 Google Dec 2025 dpaia.teamcity.com Official leaderboard
12 Junie CLI* 1512.1 + GPT-5.4 60.8 xhigh JetBrains Mar 2026 dpaia.teamcity.com Official leaderboard
13 Junie CLI* 1512.1 + GPT-5.3 Codex 60.6 xhigh JetBrains Feb 2026 dpaia.teamcity.com Official leaderboard
14 Claude Sonnet 4.6 60.1 max ultrathink, Claude Code 2.1.118 Anthropic Feb 2026 dpaia.teamcity.com Official leaderboard
15 Junie CLI* 1512.1 + Grok 4.1 Fast 47.7 reasoning JetBrains Nov 2025 dpaia.teamcity.com Official leaderboard
#ModelScoreSettingAgent byDateLinkSource
1 Claude Opus 4.7 64.9 max ultrathink, Claude Code 2.1.118 Anthropic Apr 2026 dpaia.teamcity.com Official leaderboard
2 GPT-5.4 64.7 xHigh, Codex CLI 0.124.0 OpenAI Mar 2026 dpaia.teamcity.com Official leaderboard
3 GPT-5.5 64.1 xHigh, Codex CLI 0.125.0 OpenAI Apr 2026 dpaia.teamcity.com Official leaderboard
4 Gemini 3.1 Pro 62.6 Gemini CLI 0.39.0 Google Feb 2026 dpaia.teamcity.com Official leaderboard
5 Claude Opus 4.6 62.4 max ultrathink, Claude Code 2.1.119 Anthropic Feb 2026 dpaia.teamcity.com Official leaderboard
6 Gemini 3 Flash 61.9 Gemini CLI 0.39.0 Google Dec 2025 dpaia.teamcity.com Official leaderboard
7 Claude Sonnet 4.6 60.1 max ultrathink, Claude Code 2.1.118 Anthropic Feb 2026 dpaia.teamcity.com Official leaderboard

5 models have held number one on the official leaderboard since its release in October 2025

Lab claims count from the day the card came out, independent runs from the model’s release.

Google DeepMind Anthropic OpenAI
Own leaderboard Not tested since Apr 2026

Source: Official DPAI Arena Spring Benchmark leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
60.0 61.0 62.0 63.0 64.0 65.0 2026 60.0 61.0 62.0 63.0 64.0 65.0 2026

Source: Official DPAI Arena Spring Benchmark leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Coding

All 523 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 7 months ago, in April 2026, for GPT-5.5. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (15 entries).
  • Pass Models scored 7 models we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 7 months ago, in April 2026, for GPT-5.5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.5 (#21).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on dpaia.dev: 14 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check