Model Pricing & Scores
| Model | Input $/M | Output $/M | TB | SWE-Pro | MCP-Atlas | Context | License | Status |
|---|---|---|---|---|---|---|---|---|
| Agnes 2.0 Flash | $0.10 | $0.20 $0.20 | ? | ? | ? | 256K | Proprietary | Unverified |
| Claude Fable 5 | $10 | $50 | 84.3(v) | 80(v) | 83.3 | 1M+ | Proprietary | Classifier gap |
| Claude Haiku 4.5 | $1 | $5 | ? | 39.5(s) | ? | 200K | Proprietary | Fast, unverified agent |
| Claude Mythos 5 | $10 | $50 | 88(v) | 77.8(v) | ? | 1M+ | Proprietary (restricted) | Glasswing only |
| Claude Opus 4.8 | $5 | $25 | 79~ | 69.2(v) | 77.8 | 1M | Proprietary | ★ SWE-Pro leader |
| Claude Sonnet 5 | $2 | $10 $10 | 80.4★ | 63.2 | ? | 1M | Proprietary | ★ Beats Opus on TB |
| Cohere North Mini Code | FREE | FREE | ? | ? | ? | 256K | Proprietary | Free, tool-calling native |
| DeepSeek V4 Flash | $0.14 | $0.28 | 56.9 | 52.6 | 69 | 1M | MIT | 79.0% Verified |
| DeepSeek V4 Pro | $0.44 | $0.87 | 67.9 | 76.2(a) | 73.6 | 1M | Custom | DeepSWE 8% |
| DeepSeek V4 Pro Max | $1.60 | $3.20 | ? | ? | ? | 1M | MIT | BenchLM #13 |
| ERNIE 5.1 | $0.59 | $2.65 | ? | ? | ? | 128K | Proprietary | Chinese only |
| Fugu Ultra | sub | sub | ? | 73.7(v) | ? | pool | Proprietary | Multi-agent orchestrator |
| Gemini 3.1 Pro | $2 | $12 | 54.2✗ | 46.1(s) | ? | 2M | Proprietary | 2M context, weak agent |
| Gemini 3.5 Flash | $1.50 | $9 | 76.2 | 55.1(v) | 83.6★★ | 1M+ | Proprietary | Best MCP flash |
| GLM-5.2 | $1.40 | $4.40 | 81 | 62.1(v) | 76.8 | 1M | MIT | Current main |
| GPT-5.5 | $5 | $30 | 84 | 58.6(v) | 75.3 | ~1M | Proprietary | Terminal leader |
| Grok 4.3 | $1.25-2.50 | $2.50-5 | 63.8 | ? | ? | 1M | Proprietary | BFCL 77.2 |
| Grok Build 0.1 | $1 | $2 | ? | 70.8(v) | ? | 256K | Proprietary | Fast coder + Composer |
| Kimi K2.7 Code | $0.95 | $4 | ? | 62~ | 76 | 256K | Custom | MCP-Mark 81.1 ★ |
| Llama 4 Maverick | $0.17-0.63 | $0.60-1.79 | ? | 70.4(m) | ? | 1M | Community | tau-bench 93.7 |
| MiMo v2.5 Pro | — | $3.20 | ? | ? | ? | 1M | Custom | Untested |
| MiniMax M3 | $0.30 | $1.20 $1.20 | 66(v) | 59(v) | ? | 1M | Open-weight | ★ Open TB leader |
| Mistral Large 3 | $0.50 | $1.50 | 12 | 76.2 | 70.4 | 256K | Apache 2.0 | TB 12% — avoid agents |
| Nemotron 3 Ultra | $0.37 | $1.08 | 56.4 | 71.9(v) | ? | 262K/1M | OpenMDW | 300+ tok/s |
| Qwen3-Coder-Next | ~$0.15 | ~$0.60 | ? | 70.6 | ? | 256K | Apache 2.0 | 3B active only |
| Qwen3.7 Max | $1.25 | $3.75 | 69.7 | 60.6 | 76.4 | 1M | Custom | ★ 35h proven |
| Ring 2.6 1T | $0.30 | $2.50 | ? | ? | ? | 262K | Custom | Free on OR |
| Step 3.5 Flash | $0.10 $0.10 | $0.30 $0.30 | 51(v) | 74.4(v) | ? | 256K | Apache 2.0 | ★ Cheapest coding |
| Step 3.7 Flash | $0.20 | $1.15 | 59.5 | 56.3 | ? | 256K | Apache 2.0 | Free tier + multimodal |
| Tencent Hy3 | $0.066 | $0.21 $0.21 | 54.4 | 74~ (v) | ? | 256K | Apache 2.0 | ★ Value pick |
Pricing Tiers
Budget (<$0.30/M out)
| Agnes 2.0 | $0.20 | Cheapest reasoning |
| Hy3 | $0.21 | Apache 2.0 |
| DS V4 Flash | $0.28 | Batch floor |
| Step 3.5 Flash | $0.30 | Cheapest coding/$ |
Value ($0.30-1.15/M out)
| MiniMax M3 | $1.20 | 1M ctx, open TB leader |
| Nemotron 3U | $1.08 | 300+ tok/s |
| DS V4 Pro | $0.87 | Single-pass |
| Step 3.7 | $1.15 | Free tier |
Mid ($1.20-4.40/M out)
| GLM-5.2 | $1.35-4.40 | Current main |
| Grok Build | $2 | Fast coder |
| Ring 2.6 1T | $2.50 | 1T MoE |
| Grok 4.3 | $2.50-5 | BFCL 77.2 |
| ERNIE 5.1 | $2.65 | Chinese only |
| Qwen3.7 Max | $1.29-3.75 | 35h proven |
| MiMo 2.5 Pro | $3.20 | Untested |
| Kimi K2.7 | $4 | MCP-Mark 81.1 |
| DS V4 Pro Max | $3.20 | BenchLM #13 |
Premium ($5+/M out)
| Claude Haiku 4.5 | $5 | Fast, unverified |
| Gemini 3.5 Flash | $9 | MCP leader |
| Sonnet 5 | $10 | Beats Opus TB |
| Gemini 3.1 Pro | $12 | 2M ctx |
| Opus 4.8 | $25 | Marathon 26.0 |
| GPT-5.5 | $30 | TB 84.0 |
| Fable 5 | $50 | Restored, classifier gap |