This page uses opencode/nemotron-3-super-free as the comparison baseline. Every chart and table below is intended to answer the same question: where this model leads, where it lags, and what it costs in quality, time, and request pressure.
These charts use opencode/nemotron-3-super-free as zero. Positive bars mean other models are above the baseline on that metric; negative bars mean they trail it.
Use this to decide whether another model beats opencode/nemotron-3-super-free enough to justify the change.
| Model | Composite | Delta vs baseline | Success | Success delta | ORPT | ORPT delta | Cost | Cost delta | Wall time |
|---|---|---|---|---|---|---|---|---|---|
| opencode/gpt-5.4-nano | 0.789 | +0.608 | 85% | +59% | 15.17 | -4.25 | $0.4215 | +$0.4215 | 27m 33s |
| opencode/kimi-k2.5 | 0.785 | +0.603 | 89% | +63% | 14.25 | -5.18 | $0.9122 | +$0.9122 | 41m 05s |
| opencode/claude-opus-4-6 | 0.67 | +0.489 | 89% | +63% | 14.88 | -4.55 | $21.8757 | +$21.8757 | 40m 04s |
| opencode/glm-5 | 0.623 | +0.441 | 78% | +52% | 11.57 | -7.86 | $6.4339 | +$6.4339 | 20m 10s |
| opencode/big-pickle | 0.615 | +0.434 | 67% | +41% | 15.39 | -4.04 | $0.0000 | +$0.0000 | 36m 28s |
| opencode/gpt-5.4 | 0.609 | +0.427 | 78% | +52% | 11.00 | -8.43 | $8.9827 | +$8.9827 | 32m 47s |
| opencode/claude-sonnet-4-6 | 0.593 | +0.411 | 78% | +52% | 16.43 | -3.00 | $11.8406 | +$11.8406 | 42m 31s |
| opencode/glm-5.1 | 0.547 | +0.365 | 67% | +41% | 12.06 | -7.37 | $1.8816 | +$1.8816 | 64m 39s |
| opencode/minimax-m2.5 | 0.481 | +0.299 | 56% | +30% | 18.87 | -0.56 | $0.6413 | +$0.6413 | 32m 15s |
| opencode/gpt-5.4-mini | 0.425 | +0.243 | 48% | +22% | 9.54 | -9.89 | $1.0606 | +$1.0606 | 21m 48s |
| opencode/minimax-m2.5-free | 0.415 | +0.233 | 59% | +33% | 16.19 | -3.24 | $0.0000 | +$0.0000 | 41m 34s |
| opencode/gemini-3-flash | 0.415 | +0.233 | 59% | +33% | 21.81 | +2.38 | $2.4307 | +$2.4307 | 62m 52s |
| opencode/gemini-3.1-pro | 0.291 | +0.109 | 37% | +11% | 12.70 | -6.73 | $5.8536 | +$5.8536 | 51m 25s |
| opencode/nemotron-3-super-free Baseline | 0.181 | +0.0 | 26% | +0% | 19.43 | +0.00 | $0.0000 | +$0.0000 | 109m 00s |
This table puts the most revealing tasks first: unsolved tasks, single-solver tasks, and tasks where the baseline trails the winner by a meaningful margin.
| Task | Field read | Baseline result | Winner | Gap to winner | Baseline cost | Baseline time |
|---|---|---|---|---|---|---|
| SELinux registry volume label repair | Clear separation | dnf | opencode/kimi-k2.5 1.0 |
1.0 | n/a | 5m 00s |
| RHEL k3s node preparation repair | Competitive split | failed | opencode/gpt-5.4-nano 1.0 |
1.0 | n/a | 5m 01s |
| Event status shell summary | Competitive split | dnf | opencode/big-pickle 1.0 |
1.0 | n/a | 45s |
| Kubernetes rollout repair | Clear separation | dnf | opencode/gpt-5.4-mini 1.0 |
1.0 | n/a | 5m 00s |
| Bootstrap phase validation repair | Competitive split | dnf | opencode/kimi-k2.5 0.993 |
0.993 | n/a | 5m 00s |
| ExternalDNS RFC2136 repair | Competitive split | provider-limited | opencode/kimi-k2.5 0.982 |
0.982 | n/a | 5m 01s |
| nftables router ingress repair | Competitive split | failed | opencode/gpt-5.4-nano 0.98 |
0.98 | n/a | 4m 52s |
| Docker Compose observability fix | Competitive split | failed | opencode/gpt-5.4-nano 0.975 |
0.975 | n/a | 3m 26s |
| Pre-ArgoCD bootstrap sequencing | Competitive split | dnf | opencode/gpt-5.4-nano 0.967 |
0.967 | n/a | 5m 00s |
| Log level rollup shell script | Competitive split | dnf | opencode/big-pickle 0.965 |
0.965 | n/a | 1m 00s |
| CNPG restore manifest repair | Competitive split | dnf | opencode/big-pickle 0.964 |
0.964 | n/a | 5m 00s |
| MCP OpenBao contract repair | Competitive split | dnf | opencode/big-pickle 0.954 |
0.954 | n/a | 5m 00s |
| RHEL edge firewalld router repair | Competitive split | dnf | opencode/gpt-5.4-nano 0.953 |
0.953 | n/a | 4m 00s |
| Kubernetes OIDC RBAC repair | Competitive split | dnf | opencode/gpt-5.4-nano 0.95 |
0.95 | n/a | 5m 00s |
| Log audit shell script | Competitive split | dnf | opencode/gpt-5.4-nano 0.935 |
0.935 | n/a | 1m 15s |
| Workspace runtime access convergence | Competitive split | failed | opencode/gpt-5.4-nano 0.932 |
0.932 | n/a | 5m 00s |
| Wildcard TLS route coverage | Competitive split | dnf | opencode/kimi-k2.5 0.929 |
0.929 | n/a | 5m 00s |
| MetalLB ingress address pool repair | Competitive split | failed | opencode/gpt-5.4-nano 0.928 |
0.928 | n/a | 5m 01s |
| AppArmor dnsmasq profile repair | Competitive split | failed | opencode/gpt-5.4-nano 0.918 |
0.918 | n/a | 2m 50s |
| Traefik forwarded header trust repair | Competitive split | provider-limited | opencode/kimi-k2.5 0.913 |
0.913 | n/a | 5m 01s |
| K3s registry mirror trust repair | Competitive split | passed | opencode/big-pickle 1.0 |
0.3 | n/a | 2m 01s |
| Workspace transplant bundle repair | Competitive split | passed | opencode/big-pickle 0.985 |
0.285 | n/a | 2m 40s |
| Terraform static site repair | Competitive split | passed | opencode/kimi-k2.5 0.978 |
0.278 | n/a | 3m 58s |
| Ansible nginx role completion | Competitive split | passed | opencode/big-pickle 0.963 |
0.263 | n/a | 4m 41s |
| RHEL NetworkManager bridge VLAN repair | Competitive split | passed | opencode/gpt-5.4-nano 0.951 |
0.251 | n/a | 4m 59s |
| Build workspace plane convergence | Competitive split | passed | opencode/gpt-5.4-nano 0.942 |
0.242 | n/a | 3m 23s |
| GitOps workspace render validation | Competitive split | passed | opencode/big-pickle 0.941 |
0.241 | n/a | 4m 05s |
Pairwise task wins and top-line deltas show whether a challenger truly beats the baseline or just looks cheaper or faster in isolation.
| Challenger | Task record | Composite edge | Success edge | Cost edge | Time edge | ORPT edge |
|---|---|---|---|---|---|---|
| opencode/gpt-5.4-nano | 0-23 4 ties |
-0.608 | -59% | -$0.4215 | +81m 28s | +4.25 |
| opencode/kimi-k2.5 | 1-24 2 ties |
-0.603 | -63% | -$0.9122 | +67m 56s | +5.18 |
| opencode/claude-opus-4-6 | 0-24 3 ties |
-0.489 | -63% | -$21.8757 | +68m 56s | +4.55 |
| opencode/glm-5 | 0-21 6 ties |
-0.441 | -52% | -$6.4339 | +88m 50s | +7.86 |
| opencode/big-pickle | 0-18 9 ties |
-0.434 | -41% | +$0.0000 | +72m 32s | +4.04 |
| opencode/gpt-5.4 | 0-21 6 ties |
-0.427 | -52% | -$8.9827 | +76m 13s | +8.43 |
| opencode/claude-sonnet-4-6 | 0-21 6 ties |
-0.411 | -52% | -$11.8406 | +66m 30s | +3.00 |
| opencode/glm-5.1 | 0-18 9 ties |
-0.365 | -41% | -$1.8816 | +44m 21s | +7.37 |
| opencode/minimax-m2.5 | 1-15 11 ties |
-0.299 | -30% | -$0.6413 | +76m 45s | +0.56 |
| opencode/gpt-5.4-mini | 1-13 13 ties |
-0.243 | -22% | -$1.0606 | +87m 12s | +9.89 |
| opencode/minimax-m2.5-free | 1-10 16 ties |
-0.233 | -33% | +$0.0000 | +67m 27s | +3.24 |
| opencode/gemini-3-flash | 3-12 12 ties |
-0.233 | -33% | -$2.4307 | +46m 09s | -2.38 |
| opencode/gemini-3.1-pro | 1-10 16 ties |
-0.109 | -11% | -$5.8536 | +57m 35s | +6.73 |
The benchmark result only matters in context: this section pairs the observed benchmark outcome with the catalog metadata and operating characteristics behind it.
OpenRouter reference blend for nvidia/nemotron-3-super-120b-a12b:free is 0 USD per 1M tokens using a 3:1 input:output mix. Reference price uses nvidia/nemotron-3-super-120b-a12b at 0.2 USD per 1M tokens from the same OpenRouter family.
Observed to take a slow, tool-heavy path on the scripting smoke task.