Code generation and understanding
10
Models
9
Benchmarks
5
Providers
7
Open Weight
9
Deployable Now
7 self-host
10
Structured benchmark-backed rows
0
Signal-backed rows without full normalized tables
0
Benchmark-expected rows still catching up
| # | Model | Economic Footprint | Quality | Params | SWE-Bench | TerminalBench 2.0 | GAIA | Aider Polyglot | $/M In | Speed | Open |
|---|---|---|---|---|---|---|---|---|---|---|---|
8 | CodeLlama 70B Meta Open Weights |
| 51.0 |
| 60.7 |
| 70B |
| 42.4 |
| — |
| — |
| — |
Free Free |
| — |
3 | GPT-5.1 Codex OpenAI SubscriptionStructured | 47.9 | 50.4 | Undisclosed | 66.0 | 57.8 | — | — | $1.25/M Cheapest verified | — |
14 | qwen-2.5-coder-32b-instruct Qwen Open WeightsStructured | 39.8 | 24.2 | 32B | 40.2 | — | 4.7 | 16.4 | $0.6600/M Cheapest verified | — |
4 | Qwen3 Next 80B A3B Instruct Qwen Self-HostStructured | 38.9 | 40.0 | 80B | 42.4 | — | 44.2 | 8.0 | $0.0900/M Cheapest verified | — |
1 | Grok Code Fast 1 xAI SubscriptionStructured | 37.9 | 37.0 | Undisclosed | — | 25.8 | — | — | $0.2000/M Cheapest verified | — |
6 | Qwen3 Coder 30B A3B Instruct Qwen Self-HostStructured | 36.7 | 38.4 | 30B | 51.6 | — | 44.2 | 8.0 | $0.0700/M Cheapest verified | — |
13 | Qwen3 Coder Flash Qwen Self-HostStructured | 36.4 | 31.7 | Unknown | 69.6 | — | 44.2 | 8.0 | $0.1950/M Cheapest verified | — |
7 | Devstral Medium Mistral AI Open WeightsStructured | 33.5 | 37.9 | Unknown | — | — | — | — | $0.4000/M Cheapest verified | — |
12 | Qwen3 Coder Plus Qwen Self-HostStructured | 33.5 | 31.0 | Unknown | — | — | 44.2 | — | $0.6500/M Cheapest verified | — |
10 | Codex Mini OpenAI Structured | — | — | Undisclosed | — | 43.1 | — | — | $1.50/M Cheapest verified | — |