Supabase Evals
Agents build, deploy, investigate, and repair Supabase projects. The hub port uses OpenCode and runs the original task scorers from a pinned release.
The October 28, 2026, leaderboard lists Claude Code with Opus 5.5 at 93% overall without skills. Claude Code with Sonnet 5.5 and OpenCode with Kimi K3 scored 86%.
- Made by
- Supabase
- Released
- April 2026
- Kind
- Apps and MCP servers
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Agents
- Strongest model tested
- Claude Opus 5.5 #5 on the capability index
- Last model added
- September 2026
- Scores refreshed
- November 20th, 2026
Claude Opus 5.5 leads with 92.8% among models released by September 2026
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.
The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6.1 Terra
- and 13 more
- 1 Claude Opus 5.5 Benchmark / medium 92.8%
- 2 Kimi K3 Without skills 88.4%
- 3 Claude Sonnet 5.5 Benchmark / medium 85.5%
- 4 Grok 4.7 Without skills / high 84.1%
- 5 GPT-6.1 Sol Benchmark / medium 81.2%
- 6 GPT-6 Luna Benchmark / medium 76.8%
Source: Official Supabase Evals leaderboard
| # | Model | Score | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 | 92.8% | Benchmark / medium, Claude Code | Sep 2026 | Oct 2026 | Official leaderboard |
| 2 | Kimi K3 | 88.4% | Without skills, OpenCode | Jul 2026 | Oct 2026 | Official leaderboard |
| 3 | Claude Sonnet 5.5 | 85.5% | Benchmark / medium, Claude Code | Sep 2026 | Oct 2026 | Official leaderboard |
| 4 | Grok 4.7 | 84.1% | Without skills / high, Grok Build | Sep 2026 | Oct 2026 | Official leaderboard |
| 5 | GPT-6.1 Sol | 81.2% | Benchmark / medium, Codex | Sep 2026 | Oct 2026 | Official leaderboard |
| 6 | GPT-6 Luna | 76.8% | Benchmark / medium, Codex | Sep 2026 | Oct 2026 | Official leaderboard |
Rankings by part
We read 4 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Build, Deploy and Investigate. Kimi K3 leads 3 of them, and Claude Opus 5.5 leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
Build
6 models we track have a score here.
- 1 Claude Opus 5.5 100%
- 2 Grok 4.7 ↑2 92.3%
- 3 Claude Sonnet 5.5 92.3%
- 4 Kimi K3 ↓2 87.2%
- 5 GPT-6 Luna ↑1 69.2%
- 6 GPT-6.1 Sol ↓1 69.2%
Deploy
6 models we track have a score here.
- 1 Kimi K3 ↑1 100%
- 2 Claude Opus 5.5 ↓1 100%
- 3 Claude Sonnet 5.5 100%
- 4 GPT-6.1 Sol ↑1 100%
- 5 Grok 4.7 ↓1 88.9%
- 6 GPT-6 Luna 88.9%
Investigate
6 models we track have a score here.
- 1 Kimi K3 ↑1 100%
- 2 GPT-6.1 Sol ↑3 100%
- 3 Grok 4.7 ↑1 77.8%
- 4 GPT-6 Luna ↑2 77.8%
- 5 Claude Opus 5.5 ↓4 66.7%
- 6 Claude Sonnet 5.5 ↓3 66.7%
Resolve
6 models we track have a score here.
- 1 Kimi K3 ↑1 100%
- 2 Claude Opus 5.5 ↓1 100%
- 3 GPT-6.1 Sol ↑2 100%
- 4 Grok 4.7 91.7%
- 5 GPT-6 Luna ↑1 91.7%
- 6 Claude Sonnet 5.5 ↓3 91.7%
More effort adds a median 3 points
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official Supabase Evals leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Claude Opus 5.5 | Benchmark / medium | Claude Code | 92.8% | 5,438 | |
| Claude Opus 5.5 | Without skills / medium | Claude Code | 89.9% | 5,034 | |
| Kimi K3 | Without skills | OpenCode | 88.4% | 8,950 | |
| Kimi K3 | Benchmark | OpenCode | 85.5% | 10,544 | |
| Claude Sonnet 5.5 | Benchmark / medium | Claude Code | 85.5% | 3,659 | |
| Grok 4.7 | Without skills / high | Grok Build | 84.1% | 6,861 | |
| Grok 4.7 | Benchmark / high | Grok Build | 82.6% | 6,034 | |
| GPT-6.1 Sol | Benchmark / medium | Codex | 81.2% | 6,262 | |
| GPT-6.1 Sol | Without skills / medium | Codex | 78.3% | 5,393 | |
| GPT-6 Luna | Benchmark / medium | Codex | 76.8% | 9,135 | |
| Claude Sonnet 5.5 | Without skills / medium | Claude Code | 76.8% | 2,956 | |
| GPT-6 Luna | Without skills / medium | Codex | 69.6% | 7,130 |
2 models have held number one on the official leaderboard since its release in April 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Supabase Evals leaderboard
Kimi K3 on the official leaderboard passed 85% in July 2026
That was 3 months after its release. By the same age, 23% of the benchmarks from 2026 had passed 85%. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Supabase Evals 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 Tool use
- AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state.
Gemini 4 Neon 56.3%
- MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks.
Gemini 4 Argon 90.7%
- BFCL V4 In use Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes.
Qwen3.7-Max 75%
- τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state.
Qwen 3.8 Max 55.1%
- Toolathlon-Verified In use 108 long-horizon tasks across 32 apps through about 600 MCP tools. Execution-based checkers inspect final environment state. GLM-5.3-Flash 78.4%
- DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%
Data audit
All 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (6 entries).
- Pass Models scored 6 models we track.
- Pass Newest score The newest is from September 2026, for GPT-6.1 Sol.
- Pass Frontier models 3 of the top 10 have a score, led by Claude Opus 5.5 (#5).
- Note Headroom Saturated. The best is 92.8%, past 85% since July 2026.
Leaderboard import 3 pass
- Pass Importer We read the board on supabase.com: 12 rows for models we track 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.