g
gpt-oss-120b
A single-GPU open baseline — 4.4% on APEX-Agents means simple tasks only
Intelligence v5
— v4.3.2 —
Leaderboard rank
—
Input / output
$0.05/$0.20
Blended 3:1
$0.09
Output speed
220 tok/s
Value
—
Capability profile Normalised to the field maximum
Specifications
Vendor
OpenAI
Released
2025-08
Weights
Open (Apache-2.0)
Context
131,072 tokens
Max output
Undisclosed
Parameters
117B total / 5B active
Input price
$0.05 / 1M
Output price
$0.20 / 1M
Cache hit
n/a
Output speed
220 tok/s
Intelligence
not indexed
Benchmark scores
How it is measured →Bars are scaled to the field maximum for that benchmark.
| Benchmarks | Score | Vs. best in field | Gap |
|---|---|---|---|
| AA Intelligence Index v5 | — | — | not indexed |
| AA Intelligence Index v4.3.2 | — | — | not indexed |
| GPQA Diamond Graduate-level science reasoning | 75.8 | 79% | −20.6 pts |
| Terminal-Bench 4.0 | — | — | not indexed |
| LiveBench Design | — | — | not indexed |
| Vals Index Vals composite index | 14.5 | 20% | −58.6 pts |
Assessment
117B total with 5.1B active, 131K context, Apache-2.0. Vals index 14.5 puts it last, and APEX-Agents is 4.4%. It only makes sense for one GPU, high concurrency and simple tasks. The harmony response format is mandatory — otherwise output breaks.
At the same price
The four models closest on blended price, side by side.
| Model | v5 | Blended | tok/s | Score gap |
|---|---|---|---|---|
| Claude Haiku 5.5Anthropic | 36.2 | $0.20 | 235 | — |
| MiniMax M3.1 FlashMiniMax | 29.0 | $0.26 | 118 | — |
| DeepSeek V4.1 Flash Open weightsDeepSeek | 32.5 | $0.53 | 148 | — |
| MiMo-V2.6 Pro Open weightsXiaomi | 30.0 | $0.61 | 81 | — |
Compare side by side
Line it up against other models on price, speed and every benchmark.
The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.
Open the comparison tool → Back to leaderboard